export declare const SUPPORTED_PROVIDERS: readonly ["openai", "anthropic", "google", "mistral", "xai", "meta"]; export type SupportedProvider = (typeof SUPPORTED_PROVIDERS)[number]; export type ModelType = 'chat' | 'image' | 'video' | 'audio' | 'tts' | 'stt' | 'embeddings' | 'ocr' | 'moderation' | 'responses' | 'language-completion'; export type ModelSubtype = 'reasoning' | 'code' | 'realtime'; interface BaseModelEntry { type: ModelType; provider: string; displayName: string; description?: string; subtype?: string; pricing?: { input?: number; output?: number; cachedInput?: number; longContext?: { threshold: number; input?: number; output?: number; cachedInput?: number; }; }; unitPricing?: Record; deprecated?: boolean; deprecationDate?: string; shutdownDate?: string; replacementModel?: string; isAvailable?: boolean; lastUpdated?: string; } export interface ChatModelEntry extends BaseModelEntry { type: 'chat' | 'responses' | 'language-completion'; contextLength?: number; benchmarks?: { intelligence?: number; coding?: number; math?: number; mmluPro?: number; gpqa?: number; }; performance?: { outputTokensPerSecond?: number; timeToFirstToken?: number; }; context?: { window?: number; input: { text: boolean; image: boolean; audio: boolean; video: boolean; }; output: { text: boolean; image: boolean; audio: boolean; video: boolean; }; }; sampling?: { temperature?: boolean; topP?: boolean; topK?: boolean; frequencyPenalty?: boolean; presencePenalty?: boolean; }; tools?: { supported?: boolean; }; reasoning?: { alwaysOn?: boolean; }; } export interface ImageModelEntry extends BaseModelEntry { type: 'image'; } export interface VideoModelEntry extends BaseModelEntry { type: 'video'; } export interface AudioModelEntry extends BaseModelEntry { type: 'audio'; } export interface TTSModelEntry extends BaseModelEntry { type: 'tts'; } export interface STTModelEntry extends BaseModelEntry { type: 'stt'; } export interface EmbeddingsModelEntry extends BaseModelEntry { type: 'embeddings'; contextLength?: number; } export interface OCRModelEntry extends BaseModelEntry { type: 'ocr'; } export interface ModerationModelEntry extends BaseModelEntry { type: 'moderation'; } export type ModelEntry = ChatModelEntry | ImageModelEntry | VideoModelEntry | AudioModelEntry | TTSModelEntry | STTModelEntry | EmbeddingsModelEntry | OCRModelEntry | ModerationModelEntry; export declare const MODEL_REGISTRY: { readonly 'claude-fable-5': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Fable 5"; readonly description: "Anthropic's Mythos-class model for autonomous knowledge work and coding, with 1M context and extended thinking. Exceptional performance on complex reasoning and agentic tasks. Intelligence 59.9%, coding 76.5%, GPQA 92.6%. Ideal for complex autonomous workflows and knowledge-intensive tasks."; readonly pricing: { readonly input: 10; readonly output: 50; readonly cachedInput: 1; }; readonly benchmarks: { readonly intelligence: 59.9; readonly coding: 76.5; readonly math: 99.7; readonly gpqa: 92.6; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'claude-fable-5-1': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Fable 5.1"; readonly description: "Frontier-class model with exceptional performance in software engineering, knowledge work, vision, and scientific research. Features adaptive thinking, computer use, vision, and agentic browsing/terminal. Larger lead over other models on longer and more complex tasks. 1M context window."; readonly pricing: { readonly input: 10; readonly output: 50; readonly cachedInput: 0.25; }; readonly benchmarks: { readonly intelligence: 82.95; readonly coding: 81.2; readonly math: 79.4; readonly mmluPro: 92.4; readonly gpqa: 93.4; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-12"; }; readonly 'claude-haiku-4-5-20251001': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Haiku 4.5"; readonly description: "Anthropic's fastest and most cost-efficient model. Matches Sonnet 4's performance on coding and agent tasks while maintaining near-Sonnet quality vision capabilities. Budget tier for file reads, quick edits, and routine questions. 200K context."; readonly pricing: { readonly input: 1; readonly output: 5; readonly cachedInput: 0.1; }; readonly benchmarks: { readonly coding: 73.3; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'claude-haiku-5-5': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Haiku 5.5"; readonly description: "Anthropic's fastest and most cost-efficient model. Optimized for high-volume, low-latency tasks with strong coding and reasoning capabilities. Features long-context pricing tier at 100K tokens."; readonly pricing: { readonly input: 0.1; readonly output: 0.5; readonly cachedInput: 0.01; readonly longContext: { readonly threshold: 100000; readonly input: 0.5; readonly output: 2.5; readonly cachedInput: 0.05; }; }; readonly benchmarks: { readonly intelligence: 43.4; readonly coding: 90.4; readonly gpqa: 87.1; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-23"; }; readonly 'claude-opus-4-5': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Opus 4.5"; readonly description: "Anthropic's most powerful frontier model with exceptional performance across reasoning, coding, and knowledge work. Achieves 89.5% on MMLU-Pro, 87.0% on GPQA Diamond, and record-breaking 80.9% on SWE-bench Verified. 200K context window with vision support. Ideal for complex agentic tasks and autonomous coding."; readonly pricing: { readonly input: 5; readonly output: 25; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly intelligence: 34.7; readonly coding: 80.9; readonly math: 58.6; readonly mmluPro: 89.5; readonly gpqa: 87; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-02"; }; readonly 'claude-opus-4-5-20251101': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Opus 4.5 (Nov 2025)"; readonly subtype: "reasoning"; readonly description: "Hybrid reasoning model with state-of-the-art performance on software coding tasks and agentic tasks requiring autonomous operation. 67% cheaper than Claude 3 Opus with 76% fewer output tokens. Efficiency revolution in premium intelligence."; readonly pricing: { readonly input: 5; readonly output: 25; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly coding: 79.6; readonly mmluPro: 91.6; readonly gpqa: 87; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'claude-opus-4-6': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Opus 4.6"; readonly subtype: "reasoning"; readonly description: "Anthropic's advanced reasoning model with adaptive thinking for dynamic reasoning allocation. Optimized for complex multi-step reasoning and agentic workflows. 1M context window with vision support. Competitive with frontier models on GPQA Diamond and SWE-bench benchmarks."; readonly pricing: { readonly input: 5; readonly output: 25; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly intelligence: 37.8; readonly coding: 61.3; readonly math: 59.8; readonly mmluPro: 82; readonly gpqa: 89.2; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-19"; }; readonly 'claude-opus-4-7': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Opus 4.7"; readonly subtype: "reasoning"; readonly description: "Anthropic's latest flagship model with exceptional coding and reasoning capabilities. Scores 87.6% on SWE-bench Verified and 94.2% on GPQA Diamond. Optimized for agentic tasks and complex reasoning. 1M context window with vision support."; readonly pricing: { readonly input: 5; readonly output: 25; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly intelligence: 42.7; readonly coding: 87.6; readonly math: 61.9; readonly mmluPro: 76.4; readonly gpqa: 94.2; }; readonly performance: { readonly outputTokensPerSecond: 51; readonly timeToFirstToken: 16.67; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'claude-opus-4-8': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Opus 4.8"; readonly subtype: "reasoning"; readonly description: "Anthropic's most capable model with advanced reasoning, computer use, and browser automation. Features adaptive thinking with effort control, dynamic parallel subagent workflows, and mid-task system messages. Excels at agentic tasks, code analysis, and long-context work. Scores 96.7% on USAMO 2026, 93.6% on GPQA Diamond, 88.6% on SWE-Bench Verified. 1M token context window with vision support."; readonly pricing: { readonly input: 5; readonly output: 25; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly intelligence: 96.7; readonly coding: 88.6; readonly math: 96.7; readonly gpqa: 93.6; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-21"; }; readonly 'claude-opus-5': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Opus 5"; readonly description: "Anthropic's most capable model for complex agentic coding and enterprise work. Features adaptive thinking, computer use, vision, and tool use. State-of-the-art on software coding tasks and agentic tasks requiring autonomous operation. Achieves 96% on coding benchmarks and exceptional performance on complex reasoning. 1M context window."; readonly pricing: { readonly input: 5; readonly output: 25; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly intelligence: 60.7; readonly coding: 96; readonly math: 91.3; readonly mmluPro: 91.6; readonly gpqa: 93.4; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-05"; }; readonly 'claude-opus-5-5': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Opus 5.5"; readonly description: "Anthropic's most capable model with advanced reasoning and analysis."; readonly pricing: { readonly input: 4; readonly output: 20; readonly cachedInput: 0.2; }; readonly benchmarks: { readonly coding: 54.4; readonly mmluPro: 94.3; readonly gpqa: 93.4; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-23"; }; readonly 'claude-sonnet-4-5-20250929': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Sonnet 4.5 (Sep 2025)"; readonly subtype: "reasoning"; readonly description: "Best coding model with extended thinking and hybrid reasoning. Features computer use, vision, and extended thinking capabilities. Lowest rate of successful prompt injection attacks among 23 models tested externally. 200K context."; readonly pricing: { readonly input: 3; readonly output: 15; readonly cachedInput: 0.30000000000000004; }; readonly benchmarks: { readonly coding: 79.6; readonly mmluPro: 90.4; readonly gpqa: 74.1; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-12"; }; readonly 'claude-sonnet-4-6': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Sonnet 4.6"; readonly subtype: "reasoning"; readonly description: "Most capable Sonnet model with Opus-level coding capability and 94% computer use accuracy. First Sonnet preferred over previous Opus in coding evaluations. Features full upgrade across coding, computer use, long-context reasoning, agent planning, and knowledge work. 1M context window."; readonly pricing: { readonly input: 3; readonly output: 15; readonly cachedInput: 0.3; }; readonly benchmarks: { readonly coding: 77.2; readonly mmluPro: 78; readonly gpqa: 70.8; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'claude-sonnet-5': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Sonnet 5"; readonly description: "The best combination of speed and intelligence in the Sonnet family. Most agentic Sonnet with near-Opus coding capability. Features computer use, vision, tool use, and adaptive thinking. Default model for most developers. 1M context window."; readonly pricing: { readonly input: 2; readonly output: 10; readonly cachedInput: 0.2; }; readonly benchmarks: { readonly intelligence: 53; readonly coding: 38.8; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'claude-sonnet-5-5': { readonly type: "chat"; readonly provider: "anthropic"; readonly displayName: "Claude Sonnet 5.5"; readonly description: "Anthropic's latest balanced performance model combining speed and intelligence. Strong coding, reasoning, and knowledge capabilities. 1M context window."; readonly pricing: { readonly input: 2; readonly output: 10; readonly cachedInput: 0.1; }; readonly benchmarks: { readonly intelligence: 92.1; readonly coding: 81.3; readonly math: 95.2; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'deepseek-flash': { readonly type: "chat"; readonly provider: "deepseek"; readonly displayName: "DeepSeek Flash"; readonly subtype: "reasoning"; readonly description: "DeepSeek's efficiency-optimized model with 1M context window. Supports thinking and non-thinking modes."; readonly pricing: { readonly input: 0.15; readonly output: 0.6; readonly cachedInput: 0.003; }; readonly benchmarks: { readonly coding: 49.2; readonly mmluPro: 81.2; readonly gpqa: 68.4; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-14"; }; readonly 'deepseek-v4-flash': { readonly type: "chat"; readonly provider: "deepseek"; readonly displayName: "DeepSeek V4 Flash"; readonly subtype: "reasoning"; readonly description: "Efficiency-optimized MoE model with 1M context window, supporting thinking and non-thinking modes."; readonly pricing: { readonly input: 0.22; readonly output: 0.66; readonly cachedInput: 0.007; }; readonly benchmarks: { readonly coding: 69.5; readonly mmluPro: 86.2; readonly gpqa: 88.1; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-27"; }; readonly 'deepseek-v4-flash-vision-exp': { readonly type: "chat"; readonly provider: "deepseek"; readonly displayName: "DeepSeek V4 Flash Vision (Experimental)"; readonly subtype: "reasoning"; readonly description: "Vision-capable variant of DeepSeek V4 Flash with image understanding and 1M context window."; readonly pricing: { readonly input: 0.22; readonly output: 0.66; readonly cachedInput: 0.007; }; readonly benchmarks: { readonly coding: 79; readonly mmluPro: 81.2; readonly gpqa: 88.1; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'deepseek-v4-pro': { readonly type: "chat"; readonly provider: "deepseek"; readonly displayName: "DeepSeek V4 Pro"; readonly subtype: "reasoning"; readonly description: "DeepSeek's flagship mixture-of-experts model with 1M context window. Supports both thinking and non-thinking modes for flexible reasoning. Optimized for complex reasoning and knowledge-intensive tasks. Competitive pricing at $0.66/1M input tokens."; readonly pricing: { readonly input: 0.66; readonly output: 1.98; readonly cachedInput: 0.022; }; readonly benchmarks: { readonly coding: 76.8; readonly mmluPro: 87.5; readonly gpqa: 90.1; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/deepseek-v4p1-flash': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "DeepSeek V4.1 Flash"; readonly description: "DeepSeek V4.1 Flash served via Fireworks. Fast inference variant with strong reasoning capabilities."; readonly pricing: { readonly input: 0.3; readonly output: 1.2; readonly cachedInput: 0.006; }; readonly benchmarks: { readonly coding: 74.2; readonly gpqa: 90.9; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/deepseek-v4p1-flash-1017': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "DeepSeek V4.1 Flash (1017)"; readonly description: "Dated snapshot of DeepSeek V4.1 Flash on Fireworks"; readonly pricing: { readonly input: 0.3; readonly output: 1.2; readonly cachedInput: 0.006; }; readonly contextLength: 1040000; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/ember-1': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "Ember-1"; readonly subtype: "reasoning"; readonly description: "Fireworks reasoning model built on Kimi K3 with ~40% fewer reasoning tokens"; readonly pricing: { readonly input: 3; readonly output: 15; readonly cachedInput: 0.3; }; readonly contextLength: 1040000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/glm-5p3': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "GLM 5.3"; readonly description: "Zhipu's GLM 5.3 model with strong reasoning and multilingual capabilities, served via Fireworks."; readonly pricing: { readonly input: 1.4; readonly output: 4.4; readonly cachedInput: 0.26; }; readonly benchmarks: { readonly coding: 95.4; readonly mmluPro: 86.8; readonly gpqa: 88.1; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/glm-5p3-flash': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "GLM 5.3 Flash"; readonly description: "Zhipu's GLM 5.3 Flash model optimized for speed, served via Fireworks."; readonly pricing: { readonly input: 0.15; readonly output: 0.5; readonly cachedInput: 0.03; }; readonly benchmarks: { readonly coding: 95.4; readonly mmluPro: 86.8; readonly gpqa: 88.1; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/gpt-oss-120b': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "GPT-OSS 120B"; readonly description: "OpenAI's 120B parameter open-source model with reasoning capabilities at o3/o4-mini level. Supports adjustable reasoning effort and long context."; readonly pricing: { readonly input: 0.15; readonly output: 0.6; readonly cachedInput: 0.015; }; readonly benchmarks: { readonly coding: 62.4; readonly mmluPro: 90; readonly gpqa: 68.1; }; readonly contextLength: 131072; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/inkling': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "Inkling"; readonly description: "Thinking Machines Lab's 975B parameter open-weight multimodal MoE with 41B active parameters. Supports audio and vision inputs with controllable reasoning effort."; readonly pricing: { readonly input: 1; readonly output: 4.05; }; readonly benchmarks: { readonly coding: 77.6; readonly math: 97.1; readonly mmluPro: 88.7; readonly gpqa: 87.2; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly audio: true; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-14"; }; readonly 'accounts/fireworks/models/kimi-k3': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "Kimi K3"; readonly description: "Moonshot's latest flagship model with 1M context window and multimodal capabilities. Strong performance on reasoning, coding, and long-context tasks."; readonly pricing: { readonly input: 3; readonly output: 15; readonly cachedInput: 0.3; }; readonly benchmarks: { readonly coding: 77.6; readonly math: 97.1; readonly mmluPro: 88.7; readonly gpqa: 87.2; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/minimax-m3': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "MiniMax M3"; readonly description: "MiniMax's latest multimodal model with 1M context window. Optimized for coding and agentic tasks with sparse attention."; readonly pricing: { readonly input: 0.3; readonly output: 1.2; readonly cachedInput: 0.06; }; readonly benchmarks: { readonly coding: 59; }; readonly contextLength: 512000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/muse-spark-open': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "Muse Spark Open"; readonly description: "Meta's open weights reasoning model for coding and agentic workflows, 816B total / 42B active parameters"; readonly pricing: { readonly input: 1.25; readonly output: 4.25; readonly cachedInput: 0.15; }; readonly contextLength: 1040000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/nemotron-3-ultra-nvfp4': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "Nemotron 3 Ultra (NVF4)"; readonly description: "NVIDIA's 550B parameter open-weight MoE model optimized for agentic tasks. Strong performance on reasoning, coding, and tool use."; readonly pricing: { readonly input: 0.6; readonly output: 2.4; readonly cachedInput: 0.12; }; readonly benchmarks: { readonly coding: 71.9; readonly mmluPro: 86.8; readonly gpqa: 87; }; readonly contextLength: 262144; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "Nemotron Lightning 3.5 30B A3B"; readonly description: "NVIDIA's 30B parameter MoE model (3B active) from the Nemotron-H family, built on hybrid Mamba-Transformer architecture. Configurable reasoning with multi-token prediction for low-latency serving."; readonly pricing: { readonly input: 0.05; readonly output: 0.2; readonly cachedInput: 0.01; }; readonly benchmarks: { readonly coding: 51.56; readonly mmluPro: 81.94; readonly gpqa: 75.44; }; readonly contextLength: 262144; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/models/qwen3-embedding-8b': { readonly type: "embeddings"; readonly provider: "fireworks"; readonly displayName: "Qwen 3 Embedding 8B"; readonly description: "Alibaba's 8B parameter embedding model. Efficient text embedding generation."; readonly pricing: { readonly input: 0.1; readonly output: 0; }; readonly contextLength: 40960; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly output: { readonly text: false; readonly audio: false; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-14"; }; readonly 'accounts/fireworks/models/qwen3p8-2p4t-a95b': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "Qwen3.8-2.4T-A95B"; readonly description: "Qwen 3.8 2.4T mixture-of-experts model with 95B active parameters. Optimized for agentic and coding tasks with strong multimodal capabilities. Achieves competitive performance on reasoning, coding, and knowledge benchmarks. 1M context window."; readonly pricing: { readonly input: 2; readonly output: 6; }; readonly benchmarks: { readonly coding: 85.6; readonly mmluPro: 88.6; readonly gpqa: 93.7; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'accounts/fireworks/models/qwen3p8-max': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "Qwen 3.8 Max"; readonly description: "Qwen3.8-2.4T-A95B served via Fireworks, alias for the max configuration"; readonly pricing: { readonly input: 2; readonly output: 6; readonly cachedInput: 0.25; }; readonly benchmarks: { readonly coding: 85.6; readonly mmluPro: 88.6; readonly gpqa: 93.7; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'accounts/fireworks/routers/glm-5p3-fast': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "GLM 5.3 Fast"; readonly description: "Fast router variant of Zhipu's GLM 5.3 model on Fireworks. Optimized for lower latency with premium pricing."; readonly pricing: { readonly input: 2.1; readonly output: 6.6; readonly cachedInput: 0.39; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-20"; }; readonly 'accounts/fireworks/routers/kimi-k3-fast': { readonly type: "chat"; readonly provider: "fireworks"; readonly displayName: "Kimi K3 Fast"; readonly description: "Kimi K3 fast router variant optimized for low-latency inference. Strong agentic and coding performance with competitive reasoning capabilities. Ideal for real-time applications requiring fast response times."; readonly pricing: { readonly input: 4.5; readonly output: 22.5; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'gemini-2.5-flash': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 2.5 Flash"; readonly subtype: "reasoning"; readonly description: "Google's fast and efficient model with strong performance across benchmarks. Achieves 78.3% on MMLU-Pro, 59.4% on GPQA Diamond, and 82.8% on GPQA overall. 1M context window with vision support. Optimized for speed and cost-efficiency in agentic workflows. Coding 63.9%, intelligence 14%."; readonly pricing: { readonly input: 0.3; readonly output: 2.5; readonly cachedInput: 0.03; }; readonly benchmarks: { readonly intelligence: 14; readonly coding: 63.9; readonly mmluPro: 85.9; readonly gpqa: 87.9; }; readonly performance: { readonly outputTokensPerSecond: 230.4; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-18"; }; readonly 'gemini-2.5-flash-image': { readonly type: "image"; readonly provider: "google"; readonly displayName: "Gemini 2.5 Flash Image"; readonly description: "Gemini 2.5 Flash with image output capability. Text input at $0.30/1M, image output at $30/1M tokens."; readonly pricing: { readonly input: 0.3; readonly output: 30; readonly cachedInput: 0.03; }; readonly unitPricing: { readonly per_image: 0.039; }; readonly contextLength: 32768; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-08-28"; }; readonly 'gemini-2.5-flash-lite': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 2.5 Flash Lite"; readonly subtype: "reasoning"; readonly description: "Lightweight variant of Gemini 2.5 Flash optimized for cost-efficiency and speed. Achieves 72.4% on MMLU-Pro and 47.4% on GPQA. Suitable for high-volume, latency-sensitive applications with 1M context window."; readonly pricing: { readonly input: 0.1; readonly output: 0.4; readonly cachedInput: 0.01; }; readonly benchmarks: { readonly coding: 3.7; readonly math: 92.6; readonly mmluPro: 72.4; readonly gpqa: 47.4; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-25"; }; readonly 'gemini-2.5-flash-native-audio-latest': { readonly type: "audio"; readonly provider: "google"; readonly displayName: "Gemini 2.5 Flash Native Audio Latest"; readonly description: "Latest release of Gemini 2.5 Flash Native Audio"; readonly pricing: { readonly input: 0.3; readonly output: 2.5; readonly cachedInput: 0.03; }; readonly contextLength: 131072; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-07-26"; }; readonly 'gemini-2.5-flash-native-audio-preview-12-2025': { readonly type: "audio"; readonly provider: "google"; readonly displayName: "Gemini 2.5 Flash Native Audio Preview 12-2025"; readonly description: "Gemini 2.5 Flash Native Audio Preview 12-2025"; readonly pricing: { readonly input: 0.3; readonly output: 2.5; readonly cachedInput: 0.03; }; readonly contextLength: 131072; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-07-26"; }; readonly 'gemini-2.5-flash-preview-tts': { readonly type: "tts"; readonly provider: "google"; readonly displayName: "Gemini 2.5 Flash Preview TTS"; readonly description: "Gemini 2.5 Flash text-to-speech preview. Generates natural speech from text."; readonly pricing: { readonly input: 0.5; readonly output: 10; readonly cachedInput: 0.05; }; readonly contextLength: 8192; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-16"; }; readonly 'gemini-2.5-pro': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 2.5 Pro"; readonly subtype: "reasoning"; readonly description: "Google's advanced reasoning model with exceptional performance on complex tasks. Features 1M token context window and strong multimodal capabilities. Achieves 83% on GPQA Diamond, 88.6% on MMLU-Pro, 63.8% on SWE-bench, and 18.8% on Humanity's Last Exam. Optimized for reasoning, research, and knowledge-intensive applications."; readonly pricing: { readonly input: 1.25; readonly output: 10; readonly cachedInput: 0.125; readonly longContext: { readonly threshold: 200000; readonly input: 2.5; readonly output: 15; readonly cachedInput: 0.25; }; }; readonly benchmarks: { readonly coding: 63.8; readonly math: 92; readonly mmluPro: 88.6; readonly gpqa: 83; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly audio: true; readonly image: true; readonly video: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-05"; }; readonly 'gemini-2.5-pro-preview-tts': { readonly type: "tts"; readonly provider: "google"; readonly displayName: "Gemini 2.5 Pro Preview TTS"; readonly description: "Gemini 2.5 Pro Preview TTS"; readonly pricing: { readonly input: 1; readonly output: 20; readonly cachedInput: 0.1; }; readonly contextLength: 8192; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-30"; }; readonly 'gemini-3-flash-preview': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 3 Flash Preview"; readonly subtype: "reasoning"; readonly description: "Google's high-speed thinking model designed for agentic workflows, multi-turn chat, and coding assistance. Delivers near Pro-level reasoning and tool use performance with substantially lower latency. Optimized for interactive development, long-running agent loops, and collaborative coding tasks. 1M context."; readonly pricing: { readonly input: 0.5; readonly output: 3; readonly cachedInput: 0.05; }; readonly benchmarks: { readonly coding: 76.2; readonly mmluPro: 91.8; readonly gpqa: 91.9; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-23"; }; readonly 'gemini-3-pro-image': { readonly type: "image"; readonly provider: "google"; readonly displayName: "Nano Banana Pro"; readonly description: "Gemini 3 Pro Image"; readonly pricing: { readonly input: 2; readonly output: 12; readonly cachedInput: 0.2; }; readonly unitPricing: { readonly per_image: { readonly '4k': 0.24; readonly '1k-2k': 0.134; }; }; readonly contextLength: 131072; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-21"; }; readonly 'gemini-3-pro-image-preview': { readonly type: "image"; readonly provider: "google"; readonly displayName: "Nano Banana Pro"; readonly description: "Gemini 3 Pro Image Preview"; readonly pricing: { readonly input: 2; readonly output: 120; readonly cachedInput: 0.2; }; readonly contextLength: 131072; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-09"; }; readonly 'gemini-3.1-flash-image': { readonly type: "image"; readonly provider: "google"; readonly displayName: "Nano Banana 2"; readonly description: "Gemini 3.1 Flash Image."; readonly pricing: { readonly input: 0.5; readonly output: 3; readonly cachedInput: 0.05; }; readonly unitPricing: { readonly per_image: { readonly '512': 0.045; readonly '1k': 0.067; readonly '2k': 0.101; readonly '4k': 0.15; }; }; readonly contextLength: 65536; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-21"; }; readonly 'gemini-3.1-flash-image-preview': { readonly type: "image"; readonly provider: "google"; readonly displayName: "Nano Banana 2"; readonly description: "Gemini 3.1 Flash Image Preview."; readonly pricing: { readonly input: 0.5; readonly output: 60; readonly cachedInput: 0.05; }; readonly contextLength: 65536; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'gemini-3.1-flash-lite': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 3.1 Flash Lite"; readonly subtype: "reasoning"; readonly description: "Cost-efficient 3-series model for high-volume workloads. Half the cost of Gemini 3 Flash. Strong performance with GPQA 72.2%, MMLU-Pro 83%, coding 80.6%. 1M context. Optimized for ultra-low latency and cost-effective inference."; readonly pricing: { readonly input: 0.25; readonly output: 1.5; readonly cachedInput: 0.025; }; readonly benchmarks: { readonly coding: 80.6; readonly math: 85.6; readonly mmluPro: 83; readonly gpqa: 72.2; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-20"; }; readonly 'gemini-3.1-flash-lite-image': { readonly type: "image"; readonly provider: "google"; readonly displayName: "Gemini 3.1 Flash Lite Image"; readonly description: "Google's lightweight multimodal image generation model (Nano Banana 2 Lite) based on Gemini 3.1 Flash Lite."; readonly pricing: { readonly input: 0.25; readonly output: 30; readonly cachedInput: 0.025; }; readonly contextLength: 65536; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-30"; }; readonly 'gemini-3.1-flash-lite-preview': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 3.1 Flash Lite Preview"; readonly subtype: "reasoning"; readonly description: "Preview of Gemini 3.1 Flash Lite. Cost-efficient model for high-volume tasks. GPQA 72.2%, MMLU-Pro 83%, coding 80.6%. 1M context. Optimized for ultra-low latency and cost-effective inference."; readonly pricing: { readonly input: 0.25; readonly output: 1.5; readonly cachedInput: 0.025; }; readonly benchmarks: { readonly coding: 16.2; readonly math: 85.6; readonly mmluPro: 84.5; readonly gpqa: 72.2; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-25"; }; readonly 'gemini-3.1-flash-tts-preview': { readonly type: "tts"; readonly provider: "google"; readonly displayName: "Gemini 3.1 Flash TTS Preview"; readonly description: "Gemini 3.1 Flash Lite-based text-to-speech. Cost-efficient speech generation."; readonly pricing: { readonly input: 1; readonly output: 20; readonly cachedInput: 0.1; }; readonly contextLength: 8192; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-25"; }; readonly 'gemini-3.1-pro-preview': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 3.1 Pro Preview"; readonly subtype: "reasoning"; readonly description: "Google's current flagship reasoning model with state-of-the-art performance. Achieves top GPQA score at 94.3% on science reasoning. Supports 2M context window with multimodal input (text, image, audio, video). Strong coding (80.6%), MMLU-Pro 92.6%. Ideal for complex reasoning and agentic applications. Optimized for research, knowledge work, and autonomous task execution."; readonly pricing: { readonly input: 2; readonly output: 12; readonly cachedInput: 0.2; readonly longContext: { readonly threshold: 200000; readonly input: 4; readonly output: 18; readonly cachedInput: 0.4; }; }; readonly benchmarks: { readonly intelligence: 39.5; readonly coding: 80.6; readonly math: 100; readonly mmluPro: 92.6; readonly gpqa: 94.1; }; readonly contextLength: 2000000; readonly context: { readonly input: { readonly text: true; readonly audio: true; readonly image: true; readonly video: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-20"; }; readonly 'gemini-3.1-pro-preview-customtools': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 3.1 Pro Preview Custom Tools"; readonly subtype: "reasoning"; readonly description: "Gemini 3.1 Pro variant optimized for custom tool use and function calling. GPQA 94.3%, MMLU-Pro 92.6%, coding 80.6%. 2M context."; readonly pricing: { readonly input: 2; readonly output: 12; readonly cachedInput: 0.2; readonly longContext: { readonly threshold: 200000; readonly input: 4; readonly output: 18; readonly cachedInput: 0.4; }; }; readonly benchmarks: { readonly coding: 44.7; readonly mmluPro: 95.7; readonly gpqa: 94.3; }; readonly contextLength: 2000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-25"; }; readonly 'gemini-3.5-flash': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 3.5 Flash"; readonly subtype: "reasoning"; readonly description: "Google's most intelligent Flash model, optimized for agentic workflows and coding tasks. Outperforms Gemini 3.1 Pro on challenging coding and agentic benchmarks. Delivers 42% better performance on long-range multi-turn cyber benchmarks while reducing token usage by 72%. 1M context. Coding 82.7%, MMLU-Pro 84%, GPQA 82.8%. Supports tool use and sampling control."; readonly pricing: { readonly input: 1.5; readonly output: 9; readonly cachedInput: 0.15; }; readonly benchmarks: { readonly intelligence: 55; readonly coding: 87.6; readonly mmluPro: 89.5; readonly gpqa: 92.7; }; readonly performance: { readonly outputTokensPerSecond: 280; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly audio: true; readonly image: true; readonly video: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-05"; }; readonly 'gemini-3.5-flash-lite': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 3.5 Flash-Lite"; readonly description: "Fastest and most cost-effective 3.5-class model delivering 350 output tokens per second. Optimized for low-latency and high-throughput agentic tasks with strong performance on coding, UI generation, and translation. Significantly outperforms prior Flash-Lite generations. 1M context. Supports tool use and sampling control."; readonly pricing: { readonly input: 0.3; readonly output: 2.5; readonly cachedInput: 0.03; }; readonly benchmarks: { readonly intelligence: 36; }; readonly performance: { readonly outputTokensPerSecond: 399.4; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly audio: true; readonly image: true; readonly video: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-20"; }; readonly 'gemini-3.5-live-translate-preview': { readonly type: "audio"; readonly provider: "google"; readonly displayName: "Gemini 3.5 Live Translate Preview"; readonly description: "Real-time speech-to-speech translation model supporting 70+ languages with no perceptible lag. Uses bidiGenerateContent for continuous audio streaming."; readonly pricing: { readonly input: 3.5; readonly output: 21; readonly cachedInput: 0.35000000000000003; }; readonly contextLength: 16384; readonly context: { readonly input: { readonly text: false; readonly audio: true; readonly image: false; readonly video: false; }; readonly output: { readonly text: false; readonly audio: true; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'gemini-3.6-flash': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 3.6 Flash"; readonly subtype: "reasoning"; readonly description: "Workhorse model with 17% reduction in output token usage compared to 3.5 Flash. Designed for token efficiency in coding, knowledge work, and multimodal tasks with advanced reasoning at Flash-level latency. Completes tasks 12% faster on average than predecessor. 1M context."; readonly pricing: { readonly input: 0.75; readonly output: 3.75; readonly cachedInput: 0.075; }; readonly benchmarks: { readonly intelligence: 50; }; readonly performance: { readonly outputTokensPerSecond: 231.4; readonly timeToFirstToken: 18.59; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly audio: true; readonly image: true; readonly video: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'gemini-3.7-flash': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 3.7 Flash"; readonly description: "Google's latest Gemini Flash model with enhanced capabilities. Optimized for agentic workflows and complex reasoning tasks. 1M context window with multimodal support. Supports tool use and sampling control."; readonly pricing: { readonly input: 0.75; readonly output: 3.75; readonly cachedInput: 0.075; }; readonly benchmarks: { readonly intelligence: 56; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly audio: true; readonly image: true; readonly video: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'gemini-embedding-001': { readonly type: "embeddings"; readonly provider: "google"; readonly displayName: "Gemini Embedding 001"; readonly description: "First-generation Gemini text embedding model for semantic search and RAG. Supports 2,048 token context. Ideal for cost-effective embedding applications."; readonly pricing: { readonly input: 0.15; readonly output: 0; readonly cachedInput: 0.015; }; readonly contextLength: 2048; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-20"; }; readonly 'gemini-embedding-2': { readonly type: "embeddings"; readonly provider: "google"; readonly displayName: "Gemini Embedding 2"; readonly description: "Second-generation Gemini embedding model supporting text, image, audio, and video inputs."; readonly pricing: { readonly input: 0.2; readonly output: 0; readonly cachedInput: 0.020000000000000004; }; readonly contextLength: 8192; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-16"; }; readonly 'gemini-embedding-2-preview': { readonly type: "embeddings"; readonly provider: "google"; readonly displayName: "Gemini Embedding 2 Preview"; readonly description: "Preview of Gemini Embedding 2. Multi-modal embedding model."; readonly pricing: { readonly input: 0.2; readonly output: 0; readonly cachedInput: 0.020000000000000004; }; readonly contextLength: 8192; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-07-26"; }; readonly 'gemini-flash-latest': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini Flash Latest"; readonly subtype: "reasoning"; readonly description: "Alias pointing to the latest Gemini Flash model. Fast, cost-effective multimodal model with thinking capabilities. GPQA 82.8%, MMLU 88.4%."; readonly pricing: { readonly input: 0.5; readonly output: 3; readonly cachedInput: 0.05; }; readonly benchmarks: { readonly coding: 14.1; readonly mmluPro: 55.7; readonly gpqa: 90.4; }; readonly contextLength: 1000000; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-25"; }; readonly 'gemini-flash-lite-latest': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini 3.1 Flash Lite"; readonly subtype: "reasoning"; readonly description: "Google's most cost-efficient Gemini model, optimized for ultra-low latency and high-volume, cost-sensitive applications. Ideal for translation, content moderation, UI generation, and simulations. Faster than Gemini 2.5 Flash with unmatched cost efficiency. 1M context."; readonly pricing: { readonly input: 0.25; readonly output: 1.5; readonly cachedInput: 0.025; }; readonly benchmarks: { readonly coding: 40; readonly math: 35.3; readonly mmluPro: 72.4; readonly gpqa: 47.4; }; readonly contextLength: 1000000; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-21"; }; readonly 'gemini-pro-latest': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini Pro Latest"; readonly subtype: "reasoning"; readonly description: "Alias pointing to the latest Gemini Pro model (currently Gemini 3.1 Pro). Flagship reasoning model with GPQA 94.3%, MMLU-Pro 92.6%. 2M context with multimodal capabilities."; readonly pricing: { readonly input: 2; readonly output: 12; readonly cachedInput: 0.2; }; readonly benchmarks: { readonly coding: 70.4; readonly math: 92; readonly mmluPro: 92.5; readonly gpqa: 84; }; readonly contextLength: 2000000; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-07-26"; }; readonly 'gemini-robotics-er-2-preview': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemini Robotics ER 2 Preview"; readonly subtype: "reasoning"; readonly description: "Google DeepMind's advanced robotics vision-language model for embodied AI and robotic control. Specialized for moment-finding in video (91.3% accuracy), whole-body dexterous manipulation, and gripper-based tasks. Outperforms ER 1.6 on safety and spatial reasoning. Ideal for robotics research and embodied AI applications."; readonly pricing: { readonly input: 2; readonly output: 10; readonly cachedInput: 0.2; }; readonly contextLength: 32000; readonly context: { readonly input: { readonly text: true; readonly image: true; readonly video: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'gemma-4-26b-a4b-it': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemma 4 26B A4B IT"; readonly description: "Google's 26B parameter mixture-of-experts model with 4B active parameters. Efficient inference with strong performance on reasoning and coding tasks. MMLU-Pro 82.6%, GPQA Diamond 82.3%, LiveCodeBench 77.1%. Supports text and image input."; readonly pricing: { readonly input: 0.06; readonly output: 0.3; readonly cachedInput: 0.006; }; readonly benchmarks: { readonly coding: 77.1; readonly math: 88.3; readonly mmluPro: 82.6; readonly gpqa: 82.3; }; readonly contextLength: 262144; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-12"; }; readonly 'gemma-4-31b-it': { readonly type: "chat"; readonly provider: "google"; readonly displayName: "Gemma 4 31B IT"; readonly description: "Google's 31B parameter dense multimodal model. Top-tier open-weight model with excellent performance across all benchmarks. MMLU-Pro 85.2%, GPQA Diamond 84.3%, LiveCodeBench 80%, Codeforces ELO 2150. Supports text and image input with 262K context."; readonly pricing: { readonly input: 0.12; readonly output: 0.35; readonly cachedInput: 0.012; }; readonly benchmarks: { readonly coding: 80; readonly math: 89.2; readonly mmluPro: 85.2; readonly gpqa: 84.3; }; readonly contextLength: 262144; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-12"; }; readonly 'lyria-3-clip-preview': { readonly type: "audio"; readonly provider: "google"; readonly displayName: "Lyria 3 Clip Preview"; readonly description: "Lyria 3 30s model Preview"; readonly unitPricing: { readonly per_request: 0.04; }; readonly contextLength: 1048576; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-09"; }; readonly 'lyria-3-pro-preview': { readonly type: "audio"; readonly provider: "google"; readonly displayName: "Lyria 3 Pro Preview"; readonly description: "Lyria 3 Pro music generation for full-length tracks at $0.08 per song."; readonly unitPricing: { readonly per_request: 0.08; }; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'nano-banana-pro-preview': { readonly type: "image"; readonly provider: "google"; readonly displayName: "Nano Banana Pro Preview"; readonly description: "Gemini 3 Pro Image Preview - Google's advanced image generation and editing model built on Gemini 3 Pro."; readonly pricing: { readonly input: 2; readonly output: 12; readonly cachedInput: 0.2; }; readonly unitPricing: { readonly per_image: { readonly '4k': 0.24; readonly '1k-2k': 0.134; }; }; readonly contextLength: 131072; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; readonly image: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-21"; }; readonly 'veo-3.1-fast-generate-preview': { readonly type: "video"; readonly provider: "google"; readonly displayName: "Veo 3.1 fast"; readonly description: "Veo 3.1 fast"; readonly unitPricing: { readonly per_second: { readonly '4k': 0.3; readonly '720p': 0.1; readonly '1080p': 0.12; }; }; readonly contextLength: 480; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-09"; }; readonly 'veo-3.1-generate-preview': { readonly type: "video"; readonly provider: "google"; readonly displayName: "Veo 3.1"; readonly description: "Veo 3.1"; readonly unitPricing: { readonly per_second: { readonly '4k': 0.6; readonly '720p': 0.4; readonly '1080p': 0.4; }; }; readonly contextLength: 480; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly audio: true; readonly video: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-08-28"; }; readonly 'veo-3.1-lite-generate-preview': { readonly type: "video"; readonly provider: "google"; readonly displayName: "Veo 3.1 lite"; readonly description: "Veo 3.1 lite"; readonly unitPricing: { readonly per_second: { readonly '720p': 0.05; readonly '1080p': 0.08; }; }; readonly contextLength: 480; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-12"; }; readonly 'muse-image-1.0': { readonly type: "image"; readonly provider: "meta"; readonly displayName: "Muse Image 1.0"; readonly description: "Meta's image generation and editing model. Generates and edits images from text prompts with multi-turn refinement."; readonly unitPricing: { readonly per_image: 0.01; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly image: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'muse-spark-1.1': { readonly type: "chat"; readonly provider: "meta"; readonly displayName: "Muse Spark 1.1"; readonly description: "Meta's first closed-weight multimodal reasoning model with 1M context, tool calling, structured output, and strong coding abilities."; readonly pricing: { readonly input: 1.25; readonly output: 4.25; readonly cachedInput: 0.15; }; readonly benchmarks: { readonly intelligence: 51; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly image: true; readonly video: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'muse-spark-1.2': { readonly type: "chat"; readonly provider: "meta"; readonly displayName: "Muse Spark 1.2"; readonly description: "Meta's coding-optimized multimodal reasoning model with 1M context window, built for agentic coding workloads. Features tool calling, structured output, and strong reasoning capabilities. Intelligence 54%, optimized for software engineering tasks."; readonly pricing: { readonly input: 1.25; readonly output: 4.25; readonly cachedInput: 0.15; }; readonly benchmarks: { readonly intelligence: 54; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly image: true; readonly video: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'muse-spark-1.2-contributor': { readonly type: "chat"; readonly provider: "meta"; readonly displayName: "Muse Spark 1.2 Contributor"; readonly description: "Discounted tier of Muse Spark 1.2 where Meta may use prompts and completions to improve products. Same model, lower price."; readonly pricing: { readonly input: 0.1; readonly output: 0.2; readonly cachedInput: 0.15; }; readonly benchmarks: { readonly intelligence: 54; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly image: true; readonly video: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'muse-spark-1.3': { readonly type: "chat"; readonly provider: "meta"; readonly displayName: "Muse Spark 1.3"; readonly description: "Meta's latest multimodal reasoning model with 1M context, tool calling, structured output, and strong coding abilities. Recommended default for new work."; readonly pricing: { readonly input: 1.25; readonly output: 4.25; readonly cachedInput: 0.15; }; readonly benchmarks: { readonly intelligence: 83.1; readonly coding: 75.4; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly image: true; readonly video: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'muse-spark-1.3-contributor': { readonly type: "chat"; readonly provider: "meta"; readonly displayName: "Muse Spark 1.3 Contributor"; readonly description: "Discounted tier of Muse Spark 1.3 where Meta may use prompts and completions to improve products. Same model, lower price."; readonly pricing: { readonly input: 1.25; readonly output: 4.25; readonly cachedInput: 0.15; }; readonly benchmarks: { readonly intelligence: 83.1; readonly coding: 75.4; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly image: true; readonly video: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'codestral-2508': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Codestral (2508)"; readonly subtype: "code"; readonly description: "Codestral code generation specialist. Supports FIM (Fill-In-Middle), code correction, and test generation. 256K context."; readonly pricing: { readonly input: 0.3; readonly output: 0.9; }; readonly benchmarks: { readonly coding: 86.6; }; readonly contextLength: 256000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'codestral-embed': { readonly type: "embeddings"; readonly provider: "mistral"; readonly displayName: "Codestral Embed"; readonly description: "Codestral Embed. Code-specialized embedding model."; readonly pricing: { readonly input: 0.15; readonly output: 0; }; readonly contextLength: 8192; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'codestral-embed-2505': { readonly type: "embeddings"; readonly provider: "mistral"; readonly displayName: "Codestral Embed (2505)"; readonly description: "Codestral Embed 2505. Code-specialized embedding model."; readonly pricing: { readonly input: 0.15; readonly output: 0; }; readonly contextLength: 8192; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'codestral-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Codestral Latest"; readonly subtype: "code"; readonly description: "Mistral's code-specialized model. Achieves 86.6% on HumanEval and 87.96% on HumanEvalFIM for JavaScript. Excellent for code generation and programming tasks."; readonly pricing: { readonly input: 0.3; readonly output: 0.9; }; readonly benchmarks: { readonly coding: 86.6; readonly gpqa: 44; }; readonly contextLength: 256000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'devstral-2512': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Devstral (2512)"; readonly subtype: "code"; readonly description: "Devstral 2 (2512). 123B-parameter agentic coding model. 262K context."; readonly pricing: { readonly input: 0.4; readonly output: 2; }; readonly benchmarks: { readonly coding: 46.8; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'devstral-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Devstral Latest"; readonly subtype: "code"; readonly description: "Alias for the latest Devstral model. 123B agentic coding model. 262K context."; readonly pricing: { readonly input: 0.4; readonly output: 2; }; readonly benchmarks: { readonly coding: 46.8; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'devstral-medium-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Devstral Medium Latest"; readonly subtype: "code"; readonly description: "Alias for the latest Devstral Medium model. Code generation and agentic reasoning. GPQA 49.2%. 131K context."; readonly pricing: { readonly input: 0.4; readonly output: 2; }; readonly benchmarks: { readonly coding: 3.8; readonly math: 70.7; readonly mmluPro: 49.2; readonly gpqa: 4.7; }; readonly contextLength: 131000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'devstral-small-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Devstral Small Latest"; readonly subtype: "code"; readonly description: "Mistral's specialized code generation model optimized for software engineering tasks. Compact and efficient for development workflows requiring strong coding performance. 131K context with vision support."; readonly pricing: { readonly input: 0.1; readonly output: 0.3; }; readonly benchmarks: { readonly coding: 3.7; readonly math: 63.5; readonly mmluPro: 41.4; readonly gpqa: 29.3; }; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'glm-5-2': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Z.ai GLM 5.2"; readonly description: "A third-party open source text model from Z.ai with a 1M-token context window, served via Mistral's platform."; readonly pricing: { readonly input: 1.4; readonly output: 4.4; }; readonly benchmarks: { readonly coding: 69.5; readonly math: 99.2; readonly mmluPro: 86.7; readonly gpqa: 91.2; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'magistral-medium-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Magistral Medium Latest"; readonly subtype: "reasoning"; readonly description: "Alias for the latest Magistral Medium model. Reasoning model for complex problems. 40K context."; readonly pricing: { readonly input: 2; readonly output: 5; }; readonly benchmarks: { readonly coding: 59.4; readonly gpqa: 70.8; }; readonly contextLength: 40000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'magistral-small-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Magistral Small Latest"; readonly subtype: "reasoning"; readonly description: "Alias for the latest Magistral Small model. 24B reasoning model with vision. 131K context."; readonly pricing: { readonly input: 0.5; readonly output: 1.5; }; readonly benchmarks: { readonly coding: 55.8; readonly gpqa: 68.2; }; readonly contextLength: 131000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'ministral-14b-2512': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Ministral 14B (2512)"; readonly description: "Ministral 3 14B. Largest Ministral model with vision capabilities. GPQA 57.2%. 262K context."; readonly pricing: { readonly input: 0.2; readonly output: 0.2; }; readonly benchmarks: { readonly coding: 35.1; readonly math: 30; readonly mmluPro: 69.3; readonly gpqa: 57.2; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'ministral-14b-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Ministral 14B Latest"; readonly description: "Alias for the latest Ministral 14B model. 262K context with vision. GPQA 57.2%."; readonly pricing: { readonly input: 0.2; readonly output: 0.2; }; readonly benchmarks: { readonly coding: 35.1; readonly math: 30; readonly mmluPro: 69.3; readonly gpqa: 57.2; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'ministral-3b-2512': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Ministral 3B (2512)"; readonly description: "Ministral 3 3B. Smallest Ministral model with vision capabilities. 131K context."; readonly pricing: { readonly input: 0.1; readonly output: 0.1; }; readonly benchmarks: { readonly coding: 24.7; readonly math: 22; readonly mmluPro: 52.4; readonly gpqa: 35.8; }; readonly contextLength: 131000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'ministral-3b-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Ministral 3B Latest"; readonly description: "Alias for the latest Ministral 3B model. Smallest Ministral with vision. 131K context."; readonly pricing: { readonly input: 0.1; readonly output: 0.1; }; readonly benchmarks: { readonly coding: 24.7; readonly math: 22; readonly mmluPro: 52.4; readonly gpqa: 35.8; }; readonly contextLength: 131000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'ministral-8b-2512': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Ministral 8B (2512)"; readonly description: "Ministral 3 8B. Balanced small model with vision capabilities. GPQA 47.1%. 262K context."; readonly pricing: { readonly input: 0.15; readonly output: 0.15; }; readonly benchmarks: { readonly coding: 30.3; readonly math: 31.7; readonly mmluPro: 64.2; readonly gpqa: 47.1; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'ministral-8b-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Ministral 8B Latest"; readonly description: "Alias for the latest Ministral 8B model. 262K context with vision. GPQA 47.1%."; readonly pricing: { readonly input: 0.15; readonly output: 0.15; }; readonly benchmarks: { readonly coding: 30.3; readonly math: 31.7; readonly mmluPro: 64.2; readonly gpqa: 47.1; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-code-agent-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Code Agent Latest"; readonly subtype: "code"; readonly description: "Mistral's code-specialized model for software development and autonomous coding tasks. Strong performance on coding benchmarks with 91.4% on τ³-Telecom and 77.6% on SWE-Bench Verified. Optimized for agentic coding workflows."; readonly pricing: { readonly input: 0.4; readonly output: 0.9; }; readonly benchmarks: { readonly coding: 77.6; }; readonly contextLength: 262144; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'mistral-code-fim-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Code FIM Latest"; readonly subtype: "code"; readonly description: "Mistral's Fill-In-the-Middle (FIM) code completion model. Achieves 95.3% on FIM pass@1 benchmarks, surpassing GPT-3.5 Turbo and other industry standards. Optimized for code completion and in-context code generation tasks."; readonly pricing: { readonly input: 0.3; readonly output: 0.9; }; readonly benchmarks: { readonly coding: 95.3; }; readonly contextLength: 256000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-08-19"; }; readonly 'mistral-code-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Code Latest"; readonly subtype: "code"; readonly description: "Mistral's latest code generation model. Achieves 86.6% on HumanEval and 95.3% on FIM (Fill-In-the-Middle) pass@1 tasks. 2x faster code generation than previous versions with improved tokenizer. State-of-the-art for code generation and FIM use cases."; readonly pricing: { readonly input: 0.3; readonly output: 0.9; }; readonly benchmarks: { readonly coding: 86.6; }; readonly contextLength: 256000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-08-19"; }; readonly 'mistral-embed': { readonly type: "embeddings"; readonly provider: "mistral"; readonly displayName: "Mistral Embed"; readonly description: "Mistral embedding model for semantic search and RAG."; readonly pricing: { readonly input: 0.1; readonly output: 0; }; readonly contextLength: 8192; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-embed-2312': { readonly type: "embeddings"; readonly provider: "mistral"; readonly displayName: "Mistral Embed (2312)"; readonly description: "Mistral Embed 2312. Specialized embedding model."; readonly pricing: { readonly input: 0.1; readonly output: 0; }; readonly contextLength: 8192; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-large-2512': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Large (2512)"; readonly description: "Mistral Large 3 (2512). 675B total params, 41B active via sparse MoE. Flagship model with 256K context. GPQA 48.6%, MMLU-Pro 69.7%, math 73.6%. Strong performance on reasoning and knowledge tasks."; readonly pricing: { readonly input: 0.5; readonly output: 1.5; }; readonly benchmarks: { readonly coding: 92; readonly math: 73.6; readonly mmluPro: 85.5; readonly gpqa: 43.9; }; readonly contextLength: 256000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-large-4': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Large 4"; readonly description: "Mistral's most capable model with 1M token context, reasoning, vision, and function calling. Launch discount pricing active."; readonly pricing: { readonly input: 1.36; readonly output: 4.18; readonly cachedInput: 0.14; }; readonly benchmarks: { readonly coding: 91.2; readonly math: 75.8; readonly mmluPro: 88.4; readonly gpqa: 52.1; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-large-4-0': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Large 4.0"; readonly description: "Mistral's most capable model with 1M token context, reasoning, vision, and function calling. Dated snapshot of Mistral Large 4."; readonly pricing: { readonly input: 1.36; readonly output: 4.18; readonly cachedInput: 0.14; }; readonly benchmarks: { readonly coding: 91.2; readonly math: 75.8; readonly mmluPro: 88.4; readonly gpqa: 52.1; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-large-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Large Latest"; readonly description: "Mistral's flagship reasoning model with strong multilingual capabilities. Achieves 84.7% on MMLU-Pro, 68% on GPQA Diamond, and 92% on coding benchmarks. Optimized for production workloads requiring high-quality reasoning and knowledge work. Supports 256K context window with vision and function calling."; readonly pricing: { readonly input: 0.5; readonly output: 1.5; }; readonly benchmarks: { readonly intelligence: 15.9; readonly coding: 92; readonly math: 93; readonly mmluPro: 84.7; readonly gpqa: 68; }; readonly contextLength: 256000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-medium': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Medium"; readonly subtype: "reasoning"; readonly description: "Mistral's medium-tier model for general-purpose reasoning and coding tasks. Supports multimodal inputs and function calling."; readonly pricing: { readonly input: 1.5; readonly output: 7.5; }; readonly benchmarks: { readonly coding: 65; readonly mmluPro: 72; }; readonly contextLength: 262144; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-08-19"; }; readonly 'mistral-medium-2505': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Medium 2505"; readonly description: "Versatile mid-tier model from Mistral supporting function calling and structured output. Excels at programming, mathematical reasoning, long document understanding, summarization, and dialogue. 131K token context window with cost-efficient pricing."; readonly pricing: { readonly input: 0.4; readonly output: 2; }; readonly benchmarks: { readonly coding: 4.3; readonly math: 90.7; readonly mmluPro: 57.8; readonly gpqa: 30.3; }; readonly contextLength: 131000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-medium-2508': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Medium 3 (2508)"; readonly description: "Mistral Medium 3.1. Updated enterprise model with improved multimodal support. GPQA 58.8%. 131K context."; readonly pricing: { readonly input: 0.4; readonly output: 2; }; readonly benchmarks: { readonly coding: 4.4; readonly mmluPro: 58.8; readonly gpqa: 38.3; }; readonly contextLength: 131000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-medium-2604': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Medium (2604)"; readonly subtype: "reasoning"; readonly description: "Mistral Medium 3.5. 128B instruction-following model with vision capabilities. 262K context. Strong performance on reasoning and coding tasks."; readonly pricing: { readonly input: 1.5; readonly output: 7.5; }; readonly benchmarks: { readonly coding: 4.4; readonly mmluPro: 58.8; readonly gpqa: 38.3; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-25"; }; readonly 'mistral-medium-3': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Medium 3"; readonly subtype: "reasoning"; readonly description: "Mistral's mid-tier reasoning model with strong performance on MMLU-Pro (76.0%) and GPQA (57.8%). Optimized for general-purpose tasks with 32K context window. Good balance of capability and cost for instruction-following and reasoning tasks."; readonly pricing: { readonly input: 1.5; readonly output: 7.5; }; readonly benchmarks: { readonly coding: 65; readonly mmluPro: 72; readonly gpqa: 57.8; }; readonly contextLength: 32000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-medium-3-5': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Medium 3.5"; readonly subtype: "reasoning"; readonly description: "Mistral's flagship dense 128B parameter model with 256K context window. Unified model handling instruction-following, reasoning, and coding tasks. Achieves 94.2% on τ²-Bench, 74.8% on GPQA Diamond, and 77.6% on SWE-bench Verified. Strong across all benchmark categories with vision and function calling support."; readonly pricing: { readonly input: 1.5; readonly output: 7.5; }; readonly benchmarks: { readonly coding: 77.6; readonly gpqa: 74.8; }; readonly contextLength: 256000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-medium-3.5': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Medium 3.5"; readonly subtype: "reasoning"; readonly description: "Frontier-class multimodal model optimized for agentic and coding use cases. Dense 128B parameter model with 256k context window handling instruction-following, reasoning, and coding in a single set of weights. Replaces Mistral Medium 3.1, Magistral, and Devstral 2."; readonly pricing: { readonly input: 1.5; readonly output: 7.5; }; readonly benchmarks: { readonly intelligence: 30; readonly coding: 77.6; }; readonly contextLength: 256000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-08-19"; }; readonly 'mistral-medium-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Medium Latest"; readonly description: "Mistral's balanced model optimized for cost-performance ratio. Strong performance on reasoning and knowledge tasks with 32K context window. Ideal for production applications requiring good quality at lower cost."; readonly pricing: { readonly input: 1.5; readonly output: 7.5; }; readonly benchmarks: { readonly coding: 4.4; readonly mmluPro: 85.5; readonly gpqa: 38.3; }; readonly contextLength: 131000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-moderation-2603': { readonly type: "moderation"; readonly provider: "mistral"; readonly displayName: "Mistral Moderation (2603)"; readonly description: "Mistral's latest content moderation model with improved detection capabilities. Supports classification and moderation of user-generated content."; readonly pricing: { readonly input: 0.1; }; readonly contextLength: 131072; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-ocr-2512': { readonly type: "ocr"; readonly provider: "mistral"; readonly displayName: "Mistral OCR (2512)"; readonly description: "Mistral OCR 2512. Updated document understanding model."; readonly pricing: { readonly input: 2; }; readonly unitPricing: { readonly per_page: 0.002; }; readonly contextLength: 16384; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-ocr-3': { readonly type: "ocr"; readonly provider: "mistral"; readonly displayName: "Mistral OCR 3"; readonly description: "Mistral OCR 3 advanced optical character recognition model. Achieves 74% overall win rate over OCR 2 on forms, scanned documents, complex tables, and handwriting. 88.9% accuracy on handwriting recognition. Ideal for document processing and data extraction."; readonly pricing: { readonly input: 2; }; readonly contextLength: 16384; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'mistral-ocr-3-0': { readonly type: "ocr"; readonly provider: "mistral"; readonly displayName: "Mistral OCR 3.0"; readonly description: "Mistral OCR 3.0 advanced optical character recognition model. Achieves 74% overall win rate over OCR 2 on forms, scanned documents, complex tables, and handwriting. 88.9% accuracy on handwriting recognition. Ideal for document processing and data extraction."; readonly pricing: { readonly input: 2; }; readonly contextLength: 16384; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'mistral-ocr-4': { readonly type: "ocr"; readonly provider: "mistral"; readonly displayName: "Mistral OCR 4"; readonly description: "Alias for Mistral OCR Latest. Document understanding and text extraction."; readonly pricing: { readonly input: 2; }; readonly unitPricing: { readonly per_page: 0.004; }; readonly contextLength: 16384; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-ocr-4-1': { readonly type: "ocr"; readonly provider: "mistral"; readonly displayName: "Mistral OCR 4.1"; readonly description: "Alias for Mistral OCR Latest. Latest document understanding and text extraction model with bounding boxes, block classification, and confidence scores across 170 languages."; readonly pricing: { readonly input: 2; }; readonly unitPricing: { readonly per_page: 0.004; }; readonly contextLength: 16384; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-ocr-latest': { readonly type: "ocr"; readonly provider: "mistral"; readonly displayName: "Mistral OCR Latest"; readonly description: "Mistral's state-of-the-art OCR model with 94.89% overall accuracy across diverse document types. Processes up to 2000 pages per minute. Supports 170 languages with excellent performance on scanned documents (98.96%), tables (96.12%), and math content (94.29%)."; readonly pricing: { readonly input: 2; }; readonly unitPricing: { readonly per_page: 0.004; }; readonly contextLength: 16384; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-small-2506': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Small 3.1"; readonly description: "Mistral Small 3.1 24B. Budget production model for high-volume tasks with strong performance. GPQA 50.5%, MMLU-Pro 68.1%, math 88.3%. 128K context with vision support. Excellent cost-performance ratio for enterprise deployments."; readonly pricing: { readonly input: 0.1; readonly output: 0.3; }; readonly benchmarks: { readonly coding: 27.5; readonly math: 88.3; readonly mmluPro: 68.1; readonly gpqa: 50.5; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-small-2603': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Small 4"; readonly subtype: "reasoning"; readonly description: "Mistral Small 4. Unified model combining reasoning, multimodal, and coding capabilities at 5x cheaper than competitors. GPQA 50.5%, MMLU-Pro 68.1%, math 88.3%. 262K context with vision support. Ideal for cost-sensitive reasoning applications."; readonly pricing: { readonly input: 0.15; readonly output: 0.6; }; readonly benchmarks: { readonly coding: 27.5; readonly math: 88.3; readonly mmluPro: 68.1; readonly gpqa: 50.5; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-small-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Small Latest"; readonly subtype: "reasoning"; readonly description: "Mistral's lightweight model for cost-sensitive applications. Efficient performance for general tasks."; readonly pricing: { readonly input: 0.15; readonly output: 0.6; }; readonly benchmarks: { readonly coding: 27.5; readonly math: 88.3; readonly mmluPro: 68.1; readonly gpqa: 50.5; }; readonly contextLength: 262000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'mistral-tiny-2407': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Tiny 2407"; readonly description: "Mistral's tiny and efficient model offering best-in-class text and vision capabilities. Optimized for cost-efficiency and speed. Suitable for high-volume, latency-sensitive applications."; readonly pricing: { readonly input: 0.15; readonly output: 0.15; }; readonly benchmarks: { readonly coding: 55; readonly mmluPro: 68; }; readonly contextLength: 131072; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-08-05"; }; readonly 'mistral-tiny-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Tiny Latest"; readonly description: "Mistral's efficient model for high-volume, cost-sensitive tasks. Optimized for speed and cost-effectiveness with strong reasoning capabilities. Ideal for routine applications and high-throughput inference."; readonly pricing: { readonly input: 0.15; readonly output: 0.15; }; readonly contextLength: 131072; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'mistral-vibe-cli-fast': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Vibe CLI Fast"; readonly subtype: "reasoning"; readonly description: "Fast variant of Mistral Vibe CLI for command-line interface generation and optimization. Optimized for rapid code generation with lower latency."; readonly pricing: { readonly input: 0.15; readonly output: 0.6; }; readonly benchmarks: { readonly coding: 91.2; readonly math: 75.8; readonly mmluPro: 88.4; readonly gpqa: 52.1; }; readonly contextLength: 262144; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'mistral-vibe-cli-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Vibe CLI Latest"; readonly subtype: "reasoning"; readonly description: "Latest version of Mistral Vibe CLI for command-line interface generation. Supports complex CLI tool creation and optimization."; readonly pricing: { readonly input: 1.5; readonly output: 7.5; }; readonly contextLength: 262144; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-08-19"; }; readonly 'mistral-vibe-cli-with-tools': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Mistral Vibe CLI With Tools"; readonly subtype: "reasoning"; readonly description: "Mistral's agentic coding CLI tool with tool use capabilities. Optimized for autonomous coding workflows and software development tasks. Integrates with development environments for seamless code generation and analysis."; readonly pricing: { readonly input: 1.5; readonly output: 7.5; }; readonly contextLength: 262144; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'open-mistral-nemo': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Open Mistral Nemo"; readonly description: "Mistral's 12B parameter model with 128K context. Efficient and capable for general-purpose tasks. Strong performance on multilingual benchmarks."; readonly pricing: { readonly input: 0.15; readonly output: 0.15; }; readonly benchmarks: { readonly coding: 45; readonly mmluPro: 62; readonly gpqa: 42; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'open-mistral-nemo-2407': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Open Mistral Nemo 2407"; readonly description: "Mistral Nemo model with strong performance on reasoning and knowledge tasks."; readonly pricing: { readonly input: 0.15; readonly output: 0.15; }; readonly benchmarks: { readonly coding: 50; readonly mmluPro: 62; }; readonly contextLength: 131072; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'voxtral-mini-2602': { readonly type: "audio"; readonly provider: "mistral"; readonly displayName: "Voxtral Mini Transcribe (2602)"; readonly description: "Voxtral Mini transcription. Speech-to-text at $0.003/minute."; readonly unitPricing: { readonly per_minute: 0.003; }; readonly contextLength: 16384; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'voxtral-mini-latest': { readonly type: "stt"; readonly provider: "mistral"; readonly displayName: "Voxtral Mini Latest"; readonly description: "Alias for the latest Voxtral Mini transcription model. Speech-to-text at $0.003/minute."; readonly unitPricing: { readonly per_minute: 0.003; }; readonly contextLength: 16384; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'voxtral-mini-realtime-2602': { readonly type: "audio"; readonly provider: "mistral"; readonly displayName: "Voxtral Mini Realtime (2602)"; readonly description: "Alias for Voxtral Mini Transcribe Realtime, optimized for live transcription"; readonly unitPricing: { readonly per_minute: 0.006; }; readonly contextLength: 32768; readonly context: { readonly input: { readonly text: false; readonly audio: true; readonly image: false; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-27"; }; readonly 'voxtral-mini-realtime-latest': { readonly type: "audio"; readonly provider: "mistral"; readonly displayName: "Voxtral Mini Realtime Latest"; readonly description: "Latest alias for Voxtral Mini Transcribe Realtime model"; readonly unitPricing: { readonly per_minute: 0.006; }; readonly contextLength: 32768; readonly context: { readonly input: { readonly text: false; readonly audio: true; readonly image: false; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'voxtral-mini-transcribe-realtime-2602': { readonly type: "audio"; readonly provider: "mistral"; readonly displayName: "Voxtral Mini Transcribe Realtime"; readonly description: "Mistral's realtime transcription model with sub-200ms latency at $0.006/min"; readonly unitPricing: { readonly per_minute: 0.006; }; readonly contextLength: 32768; readonly context: { readonly input: { readonly text: false; readonly audio: true; readonly image: false; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'voxtral-mini-tts-2603': { readonly type: "tts"; readonly provider: "mistral"; readonly displayName: "Voxtral Mini TTS (2603)"; readonly description: "Voxtral Mini TTS. Text-to-speech with zero-shot voice cloning and multilingual support."; readonly unitPricing: { readonly per_1m_characters: 16; }; readonly contextLength: 4096; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'voxtral-mini-tts-latest': { readonly type: "tts"; readonly provider: "mistral"; readonly displayName: "Voxtral Mini TTS Latest"; readonly description: "Alias for the latest Voxtral Mini TTS model."; readonly unitPricing: { readonly per_1m_characters: 16; }; readonly contextLength: 4096; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'voxtral-small-2507': { readonly type: "stt"; readonly provider: "mistral"; readonly displayName: "Voxtral Small (2507)"; readonly description: "Voxtral Small 24B. Audio-capable model with $100/1M seconds audio input pricing. 32K context."; readonly pricing: { readonly input: 0.1; readonly output: 0.4; }; readonly unitPricing: { readonly per_minute: 0.004; }; readonly contextLength: 32000; readonly context: { readonly input: { readonly text: true; readonly audio: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'voxtral-small-latest': { readonly type: "stt"; readonly provider: "mistral"; readonly displayName: "Voxtral Small Latest"; readonly description: "Alias for the latest Voxtral Small model. Audio-capable 24B model. 32K context."; readonly pricing: { readonly input: 0.1; readonly output: 0.4; }; readonly unitPricing: { readonly per_minute: 0.004; }; readonly contextLength: 32000; readonly context: { readonly input: { readonly text: true; readonly audio: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'zai-glm-5': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Z.ai GLM 5"; readonly description: "Official Z.ai GLM 5.3 model served via Mistral platform. 1M token context with reasoning and function calling."; readonly pricing: { readonly input: 1.4; readonly output: 4.4; readonly cachedInput: 0.14; }; readonly benchmarks: { readonly coding: 95.4; readonly mmluPro: 86.8; readonly gpqa: 88.1; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'zai-glm-5-2': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Z.ai GLM 5.2"; readonly description: "GLM-5.2 753B parameter mixture-of-experts model with 1M context window. Strong performance on reasoning and coding benchmarks. Leads on SWE-bench Pro and MCP-Atlas. Optimized for complex reasoning and agentic tasks. MIT open-source license."; readonly pricing: { readonly input: 1.4; readonly output: 4.4; readonly cachedInput: 0.14; }; readonly benchmarks: { readonly intelligence: 33.7; readonly coding: 68.8; readonly gpqa: 89.5; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'zai-glm-5-3': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Z.ai GLM 5.3"; readonly description: "A third-party open weight text model from Z.ai with a 1M-token context window, served via Mistral's platform."; readonly pricing: { readonly input: 1.4; readonly output: 4.4; readonly cachedInput: 0.14; }; readonly benchmarks: { readonly intelligence: 44.8; readonly coding: 74.8; readonly gpqa: 91.7; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'zai-glm-latest': { readonly type: "chat"; readonly provider: "mistral"; readonly displayName: "Z.ai GLM Latest"; readonly description: "Alias for the latest Z.ai GLM model served via Mistral's platform."; readonly pricing: { readonly input: 1.4; readonly output: 4.4; readonly cachedInput: 0.14; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'chat-latest': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "Chat Latest (GPT-5.5 Instant)"; readonly description: "Alias for GPT-5.5 Instant. Points to the latest fast chat model. 1M context with vision."; readonly pricing: { readonly input: 5; readonly output: 30; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly intelligence: 60; readonly gpqa: 93.5; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'chatgpt-image-latest': { readonly type: "image"; readonly provider: "openai"; readonly displayName: "ChatGPT Image Latest"; readonly description: "ChatGPT image generation model. Token-based pricing: $5/1M text input, $8/1M image input, $30/1M output."; readonly pricing: { readonly input: 5; readonly output: 30; readonly cachedInput: 0.5; }; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly image: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-30"; }; readonly 'gpt-3.5-turbo': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-3.5 Turbo"; readonly description: "Legacy fast affordable model. 16K context. Superseded by GPT-4o Mini for most use cases. Coding 10.7%, math 44.1%, MMLU-Pro 46.2%, GPQA 29.7%. Still useful for cost-sensitive applications with basic requirements."; readonly pricing: { readonly input: 0.5; readonly output: 1.5; }; readonly benchmarks: { readonly coding: 10.7; readonly math: 44.1; readonly mmluPro: 46.2; readonly gpqa: 29.7; }; readonly contextLength: 16385; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-3.5-turbo-0125': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-3.5 Turbo (0125)"; readonly description: "GPT-3.5 Turbo January 2025 snapshot"; readonly pricing: { readonly input: 0.5; readonly output: 1.5; }; readonly benchmarks: { readonly coding: 10.7; readonly math: 44.1; readonly mmluPro: 46.2; readonly gpqa: 29.7; }; readonly contextLength: 16385; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-3.5-turbo-16k': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-3.5 Turbo 16K"; readonly description: "GPT-3.5 Turbo with 16K context"; readonly pricing: { readonly input: 0.5; readonly output: 1.5; }; readonly benchmarks: { readonly coding: 10.7; readonly math: 44.1; readonly mmluPro: 46.2; readonly gpqa: 29.7; }; readonly contextLength: 16385; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-13"; }; readonly 'gpt-4': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4"; readonly description: "OpenAI's GPT-4 model, a large multimodal model capable of processing both text and image inputs. 8,192 token context window. Coding 13.1%, math 64.5%, MMLU-Pro 89.2%, GPQA 53.6%. Strong performance on complex reasoning tasks."; readonly pricing: { readonly input: 30; readonly output: 60; }; readonly benchmarks: { readonly coding: 13.1; readonly math: 64.5; readonly mmluPro: 89.2; readonly gpqa: 53.6; }; readonly contextLength: 8192; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-27"; }; readonly 'gpt-4-0613': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4 (0613)"; readonly description: "OpenAI's GPT-4 model snapshot from June 13, 2023. Supports text and image inputs with 8,192 token context window. Coding 13.1%, math 64.5%, MMLU-Pro 89.2%, GPQA 53.6%. Strong performance on complex reasoning."; readonly pricing: { readonly input: 30; readonly output: 60; }; readonly benchmarks: { readonly coding: 13.1; readonly math: 64.5; readonly mmluPro: 89.2; readonly gpqa: 53.6; }; readonly contextLength: 8192; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-4-turbo': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4 Turbo"; readonly description: "OpenAI's enhanced version of GPT-4 with significantly increased context window and access to more up-to-date knowledge. 128K context window. Delivers greatly reduced response times compared to standard GPT-4. Coding 29.1%, math 73.7%, MMLU-Pro 69.4%, GPQA 66.3%. Ideal for applications requiring instant feedback and complex reasoning."; readonly pricing: { readonly input: 5; readonly output: 15; }; readonly benchmarks: { readonly coding: 29.1; readonly math: 73.7; readonly mmluPro: 69.4; readonly gpqa: 66.3; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-27"; }; readonly 'gpt-4-turbo-2024-04-09': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4 Turbo (2024-04-09)"; readonly description: "GPT-4 Turbo April 2024 snapshot with 128K context and vision. GPQA 66.3%, MMLU-Pro 69.4%, math 73.7%, coding 94.2%. Strong reasoning and coding performance."; readonly pricing: { readonly input: 10; readonly output: 30; }; readonly benchmarks: { readonly coding: 94.2; readonly math: 73.7; readonly mmluPro: 86.5; readonly gpqa: 48; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-4.1': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4.1"; readonly description: "OpenAI's capable model with 1M token context window. Exceptional performance across benchmarks with MMLU-Pro 90.2%, GPQA 66.3%, and strong coding (51.6%). Ideal for complex reasoning and knowledge-intensive tasks with extended context requirements. Supports tool use and sampling control."; readonly pricing: { readonly input: 2; readonly output: 8; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly intelligence: 38; readonly coding: 51.6; readonly math: 48.1; readonly mmluPro: 90.2; readonly gpqa: 66.3; }; readonly contextLength: 1047576; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-4.1-2025-04-14': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4.1 (2025-04-14)"; readonly description: "GPT-4.1 snapshot from April 2025. High-capability model with 1M context window. Strong performance on MMLU-Pro (90.2%) and GPQA (66.3%). Excellent for complex reasoning and knowledge tasks."; readonly pricing: { readonly input: 2; readonly output: 8; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly coding: 54.6; readonly math: 91.3; readonly mmluPro: 90.2; readonly gpqa: 66.3; }; readonly contextLength: 1047576; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-4.1-mini': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4.1 Mini"; readonly description: "OpenAI's compact model with 1M context. Excellent cost-performance ratio for coding and general tasks. MMLU-Pro 87.5%, GPQA 65%, strong instruction-following. Ideal for high-volume applications requiring strong reasoning at lower cost. Supports tool use and sampling control."; readonly pricing: { readonly input: 0.4; readonly output: 1.6; readonly cachedInput: 0.1; }; readonly benchmarks: { readonly intelligence: 34; readonly coding: 34.7; readonly math: 49.6; readonly mmluPro: 87.5; readonly gpqa: 65; }; readonly contextLength: 1047576; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-23"; }; readonly 'gpt-4.1-mini-2025-04-14': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4.1 Mini (2025-04-14)"; readonly description: "GPT-4.1 Mini snapshot from April 2025. Compact model with 1M context. MMLU-Pro 90.2%, GPQA 66.3%, strong instruction-following. Ideal for cost-effective applications."; readonly pricing: { readonly input: 0.4; readonly output: 1.6; readonly cachedInput: 0.1; }; readonly benchmarks: { readonly coding: 54.6; readonly math: 92.5; readonly mmluPro: 90.2; readonly gpqa: 66.3; }; readonly contextLength: 1047576; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-23"; }; readonly 'gpt-4.1-nano': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4.1 Nano"; readonly description: "OpenAI's smallest model with 1M context. Ultra-low cost at $0.10/$0.40 per 1M tokens. Best for high-volume simple tasks with MMLU-Pro 80.1% and GPQA 50.3%. Ideal for cost-sensitive applications with basic reasoning requirements. Supports tool use and sampling control."; readonly pricing: { readonly input: 0.1; readonly output: 0.4; readonly cachedInput: 0.025; }; readonly benchmarks: { readonly intelligence: 26; readonly coding: 9.8; readonly math: 29.4; readonly mmluPro: 80.1; readonly gpqa: 50.3; }; readonly contextLength: 1047576; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-23"; }; readonly 'gpt-4.1-nano-2025-04-14': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4.1 Nano (2025-04-14)"; readonly description: "GPT-4.1 Nano snapshot from April 2025. Ultra-efficient model with 1M context. MMLU-Pro 90.2%, GPQA 66.3%. Ideal for cost-sensitive applications."; readonly pricing: { readonly input: 0.1; readonly output: 0.4; readonly cachedInput: 0.025; }; readonly benchmarks: { readonly coding: 54.6; readonly math: 84.8; readonly mmluPro: 90.2; readonly gpqa: 66.3; }; readonly contextLength: 1047576; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-23"; }; readonly 'gpt-4o': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4o"; readonly description: "OpenAI's flagship multimodal model with fast inference and strong performance across reasoning, coding, and knowledge tasks. Optimized for speed and cost-efficiency. 128K context window with vision and audio input support. MMLU-Pro 85.7%, GPQA 54.3%, coding 33.3%, math 34.5%. Excellent for production applications requiring balanced performance and latency. Supports tool use and sampling control."; readonly pricing: { readonly input: 2.5; readonly output: 10; readonly cachedInput: 1.25; }; readonly benchmarks: { readonly intelligence: 11.2; readonly coding: 33.3; readonly math: 34.5; readonly mmluPro: 85.7; readonly gpqa: 54.3; }; readonly performance: { readonly outputTokensPerSecond: 63; readonly timeToFirstToken: 53.76; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: true; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-4o-2024-05-13': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4o (2024-05-13)"; readonly description: "GPT-4o initial release snapshot. Multimodal model with 128K context. Coding 90.2%, math 76.6%, MMLU-Pro 88.7%, GPQA 53.6%. Foundation model for GPT-4o series."; readonly pricing: { readonly input: 5; readonly output: 15; readonly cachedInput: 2.5; }; readonly benchmarks: { readonly coding: 90.2; readonly math: 76.6; readonly mmluPro: 88.7; readonly gpqa: 53.6; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-4o-2024-08-06': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4o (2024-08-06)"; readonly description: "GPT-4o August 2024 snapshot. Multimodal model with 128K context and vision support. Coding 90.2%, math 76.6%, MMLU-Pro 88.7%, GPQA 53.6%. Strong all-around performance."; readonly pricing: { readonly input: 2.5; readonly output: 10; readonly cachedInput: 1.25; }; readonly benchmarks: { readonly coding: 90.2; readonly math: 76.6; readonly mmluPro: 88.7; readonly gpqa: 53.6; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-4o-2024-11-20': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4o (2024-11-20)"; readonly description: "GPT-4o November 2024 snapshot. Multimodal model with 128K context. Strong performance across reasoning and coding. MMLU-Pro 88.7%, GPQA 53.6%, coding 90.2%. Ideal for production applications."; readonly pricing: { readonly input: 2.5; readonly output: 10; readonly cachedInput: 1.25; }; readonly benchmarks: { readonly coding: 90.2; readonly math: 76.6; readonly mmluPro: 88.7; readonly gpqa: 53.6; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-23"; }; readonly 'gpt-4o-mini': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4o Mini"; readonly description: "Compact multimodal model delivering strong reasoning performance for its size. Outperforms competitors on MMLU while maintaining low latency and cost. 128K context window with vision support. MMLU-Pro 41.5%, GPQA 42.6%, coding 34.9%. Ideal for high-volume applications requiring fast, cost-effective inference."; readonly pricing: { readonly input: 0.15; readonly output: 0.6; readonly cachedInput: 0.075; }; readonly benchmarks: { readonly intelligence: 6.9; readonly coding: 87.2; readonly math: 87; readonly mmluPro: 82; readonly gpqa: 42.6; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-21"; }; readonly 'gpt-4o-mini-2024-07-18': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-4o Mini (2024-07-18)"; readonly description: "GPT-4o Mini snapshot from July 2024 with 128K context. Compact multimodal model with strong reasoning. GPQA 42.6%, MMLU-Pro 82%. Ideal for cost-effective applications requiring vision support."; readonly pricing: { readonly input: 0.15; readonly output: 0.6; readonly cachedInput: 0.075; }; readonly benchmarks: { readonly coding: 4; readonly math: 78.9; readonly mmluPro: 14.7; readonly gpqa: 50.3; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-4o-mini-transcribe': { readonly type: "audio"; readonly provider: "openai"; readonly displayName: "GPT-4o Mini Transcribe"; readonly description: "GPT-4o Mini speech-to-text transcription"; readonly pricing: { readonly input: 1.25; readonly output: 5; readonly cachedInput: 0.625; }; readonly unitPricing: { readonly per_minute: 0.003; }; readonly context: { readonly input: { readonly audio: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-4o-mini-transcribe-2025-03-20': { readonly type: "audio"; readonly provider: "openai"; readonly displayName: "GPT-4o Mini Transcribe (2025-03-20)"; readonly description: "GPT-4o Mini Transcribe March 2025 snapshot"; readonly pricing: { readonly input: 1.25; readonly output: 5; readonly cachedInput: 0.625; }; readonly unitPricing: { readonly per_minute: 0.003; }; readonly context: { readonly input: { readonly audio: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-4o-mini-transcribe-2025-12-15': { readonly type: "stt"; readonly provider: "openai"; readonly displayName: "GPT-4o Mini Transcribe (2025-12-15)"; readonly description: "Dated snapshot of GPT-4o Mini Transcribe"; readonly pricing: { readonly input: 1.25; readonly output: 5; readonly cachedInput: 0.625; }; readonly unitPricing: { readonly per_minute: 0.003; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; }; readonly sampling: { readonly topP: true; readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-4o-mini-tts': { readonly type: "tts"; readonly provider: "openai"; readonly displayName: "GPT-4o Mini TTS"; readonly description: "Lightweight text-to-speech model based on GPT-4o Mini. Optimized for cost-effective audio generation. Supports natural-sounding speech synthesis at $15 per 1M characters. Ideal for high-volume TTS applications."; readonly unitPricing: { readonly per_1m_characters: 15; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-25"; }; readonly 'gpt-4o-mini-tts-2025-03-20': { readonly type: "tts"; readonly provider: "openai"; readonly displayName: "GPT-4o Mini TTS (2025-03-20)"; readonly description: "GPT-4o Mini TTS March 2025 snapshot. Lightweight text-to-speech model optimized for cost-effective audio generation. Supports natural-sounding speech synthesis."; readonly unitPricing: { readonly per_1m_characters: 15; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-25"; }; readonly 'gpt-4o-mini-tts-2025-12-15': { readonly type: "tts"; readonly provider: "openai"; readonly displayName: "GPT-4o Mini TTS (2025-12-15)"; readonly description: "GPT-4o Mini TTS December 2025 snapshot"; readonly unitPricing: { readonly per_1m_characters: 15; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-21"; }; readonly 'gpt-4o-transcribe': { readonly type: "stt"; readonly provider: "openai"; readonly displayName: "GPT-4o Transcribe"; readonly description: "GPT-4o speech-to-text transcription"; readonly pricing: { readonly input: 2.5; readonly output: 10; readonly cachedInput: 1.25; }; readonly unitPricing: { readonly per_minute: 0.006; }; readonly context: { readonly input: { readonly audio: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-4o-transcribe-diarize': { readonly type: "stt"; readonly provider: "openai"; readonly displayName: "GPT-4o Transcribe Diarize"; readonly description: "GPT-4o transcription with speaker diarization"; readonly pricing: { readonly input: 2.5; readonly output: 10; readonly cachedInput: 1.25; }; readonly unitPricing: { readonly per_minute: 0.006; }; readonly context: { readonly input: { readonly audio: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-5-nano': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5 Nano"; readonly subtype: "reasoning"; readonly description: "OpenAI's smallest and cheapest model. Efficient reasoning at ultra-low cost. 128K context with vision. GPQA 67.6%, MMLU-Pro 82%. Ideal for high-volume reasoning tasks. Supports tool use and sampling control."; readonly pricing: { readonly input: 0.05; readonly output: 0.4; readonly cachedInput: 0.005; }; readonly benchmarks: { readonly coding: 72.8; readonly math: 98.9; readonly mmluPro: 82; readonly gpqa: 67.6; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-5-nano-2025-08-07': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5 Nano (2025-08-07)"; readonly subtype: "reasoning"; readonly description: "GPT-5 Nano snapshot from August 2025. Compact reasoning model. 128K context. GPQA 67.6%, MMLU-Pro 82%. Strong performance on reasoning tasks at minimal cost."; readonly pricing: { readonly input: 0.05; readonly output: 0.4; readonly cachedInput: 0.005; }; readonly benchmarks: { readonly coding: 72.8; readonly math: 98.9; readonly mmluPro: 82; readonly gpqa: 67.6; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-5-search-api': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5 Search API"; readonly description: "GPT-5 with web search capabilities. 128K context. Combines frontier reasoning with real-time web search for knowledge-intensive tasks. GPQA 68.6%, MMLU-Pro 82%. Ideal for research and information-gathering applications."; readonly pricing: { readonly input: 1.25; readonly output: 10; readonly cachedInput: 0.125; }; readonly benchmarks: { readonly coding: 5.8; readonly mmluPro: 48.3; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-5-search-api-2025-10-14': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5 Search API (2025-10-14)"; readonly description: "GPT-5 Search API snapshot from October 2025. Combines frontier reasoning with web search. 128K context. GPQA 68.6%, MMLU-Pro 82%. Ideal for knowledge-intensive tasks requiring current information."; readonly pricing: { readonly input: 1.25; readonly output: 10; readonly cachedInput: 0.125; }; readonly benchmarks: { readonly coding: 5.8; readonly mmluPro: 48.3; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-5.1': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.1"; readonly subtype: "reasoning"; readonly description: "OpenAI's high-capability reasoning model with 128K context. Strong performance across coding (76.3%), math, and reasoning benchmarks. MMLU-Pro 90.2%, GPQA 88.1%. Excellent for complex reasoning, code generation, and knowledge-intensive tasks."; readonly pricing: { readonly input: 1.25; readonly output: 10; readonly cachedInput: 0.125; }; readonly benchmarks: { readonly coding: 76.3; readonly mmluPro: 90.2; readonly gpqa: 88.1; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-5.1-2025-11-13': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.1 (2025-11-13)"; readonly subtype: "reasoning"; readonly description: "GPT-5.1 snapshot from November 2025"; readonly pricing: { readonly input: 1.25; readonly output: 10; readonly cachedInput: 0.125; }; readonly benchmarks: { readonly coding: 76.3; readonly gpqa: 88.1; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-23"; }; readonly 'gpt-5.2': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.2"; readonly subtype: "reasoning"; readonly description: "OpenAI's flagship reasoning model with exceptional math performance (100%) and strong coding (55.6%). 128K context window with vision support. MMLU-Pro 81.4%, GPQA 92.4%. State-of-the-art on mathematical reasoning and complex problem-solving. Ideal for STEM applications and research."; readonly pricing: { readonly input: 1.75; readonly output: 14; readonly cachedInput: 0.175; }; readonly benchmarks: { readonly coding: 55.6; readonly math: 100; readonly mmluPro: 81.4; readonly gpqa: 92.4; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-5.2-2025-12-11': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.2 (2025-12-11)"; readonly subtype: "reasoning"; readonly description: "GPT-5.2 date-stamped release. Flagship reasoning model with exceptional math performance (100%) and strong coding (55.6%). 128K context. MMLU-Pro 90.2%, GPQA 92.4%. Ideal for mathematical reasoning and STEM applications."; readonly pricing: { readonly input: 1.75; readonly output: 14; readonly cachedInput: 0.175; }; readonly benchmarks: { readonly coding: 55.6; readonly math: 100; readonly gpqa: 92.4; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-5.4': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.4"; readonly subtype: "reasoning"; readonly description: "OpenAI's latest flagship reasoning model with state-of-the-art performance. GPQA Diamond 92.8%, MMLU-Pro 94%, math 94%. Codeforces state-of-the-art. 128K context window. Ideal for complex reasoning, mathematics, and competitive programming tasks."; readonly pricing: { readonly input: 2.5; readonly output: 15; readonly cachedInput: 0.25; readonly longContext: { readonly threshold: 272000; readonly input: 5; readonly output: 22.5; readonly cachedInput: 0.5; }; }; readonly benchmarks: { readonly intelligence: 57; readonly coding: 75.1; readonly math: 82.5; readonly mmluPro: 94; readonly gpqa: 93.3; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-5.4-2026-03-05': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.4 (2026-03-05)"; readonly subtype: "reasoning"; readonly description: "GPT-5.4 dated snapshot from March 2026. State-of-the-art reasoning model. GPQA Diamond 92.8%, MMLU-Pro 94%, strong across all benchmarks. 128K context. Ideal for complex reasoning and knowledge-intensive tasks."; readonly pricing: { readonly input: 2.5; readonly output: 15; readonly cachedInput: 0.25; readonly longContext: { readonly threshold: 272000; readonly input: 5; readonly output: 22.5; readonly cachedInput: 0.5; }; }; readonly benchmarks: { readonly coding: 88; readonly math: 94; readonly mmluPro: 94; readonly gpqa: 92.8; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-5.4-mini': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.4 Mini"; readonly subtype: "reasoning"; readonly description: "Smaller, faster, cheaper GPT-5.4 variant with strong reasoning. 128K context with vision. Excellent cost-performance ratio for reasoning tasks. AIME 98.9%, MMLU-Pro 90%, GPQA 85%. Ideal for cost-conscious applications requiring advanced reasoning."; readonly pricing: { readonly input: 0.75; readonly output: 4.5; readonly cachedInput: 0.075; }; readonly benchmarks: { readonly coding: 72.8; readonly math: 98.9; readonly mmluPro: 90; readonly gpqa: 85; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-5.4-mini-2026-03-17': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.4 Mini (2026-03-17)"; readonly subtype: "reasoning"; readonly description: "GPT-5.4 Mini dated snapshot from March 2026. Compact reasoning model with 128K context. Strong performance on AIME (98.9%) and reasoning tasks. MMLU-Pro 90%, GPQA 85%."; readonly pricing: { readonly input: 0.75; readonly output: 4.5; readonly cachedInput: 0.075; }; readonly benchmarks: { readonly coding: 72.8; readonly math: 98.9; readonly mmluPro: 90; readonly gpqa: 85; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-5.4-nano': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.4 Nano"; readonly subtype: "reasoning"; readonly description: "Smallest and cheapest GPT-5.4 variant. Efficient reasoning at ultra-low cost. 128K context with vision. AIME 98.9%, MMLU-Pro 83.2%, GPQA 77.6%. Ideal for high-volume reasoning tasks with minimal cost."; readonly pricing: { readonly input: 0.2; readonly output: 1.25; readonly cachedInput: 0.02; }; readonly benchmarks: { readonly coding: 72.8; readonly math: 98.9; readonly mmluPro: 83.2; readonly gpqa: 77.6; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-5.4-nano-2026-03-17': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.4 Nano (2026-03-17)"; readonly subtype: "reasoning"; readonly description: "GPT-5.4 Nano dated snapshot from March 2026. Ultra-efficient reasoning model. 128K context. AIME 98.9%, MMLU-Pro 83.2%, GPQA 77.6%. Ideal for cost-sensitive reasoning applications."; readonly pricing: { readonly input: 0.2; readonly output: 1.25; readonly cachedInput: 0.02; }; readonly benchmarks: { readonly coding: 72.8; readonly math: 98.9; readonly mmluPro: 83.2; readonly gpqa: 77.6; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'gpt-5.5': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.5"; readonly description: "OpenAI's smartest model (April 2026). Excels at agentic coding, computer use, knowledge work, and scientific research. GPQA 93.5%, SWE-bench Verified 80.6%, Terminal-Bench 84.7%. 1M context window with vision support. Optimized for complex multi-step reasoning and autonomous task execution. Supports tool use and sampling control."; readonly pricing: { readonly input: 5; readonly output: 30; readonly cachedInput: 0.5; readonly longContext: { readonly threshold: 272000; readonly input: 10; readonly output: 45; readonly cachedInput: 1; }; }; readonly benchmarks: { readonly intelligence: 60; readonly coding: 85.3; readonly math: 100; readonly mmluPro: 88.1; readonly gpqa: 93.6; }; readonly performance: { readonly outputTokensPerSecond: 63; readonly timeToFirstToken: 53.76; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-5.5-2026-04-23': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.5 (2026-04-23)"; readonly description: "Dated snapshot of GPT-5.5 from April 23, 2026. Frontier-class model with exceptional reasoning and coding. GPQA 93.5%, coding 80.6%. 1M context window. Ideal for complex autonomous workflows."; readonly pricing: { readonly input: 5; readonly output: 30; readonly cachedInput: 0.5; readonly longContext: { readonly threshold: 272000; readonly input: 10; readonly output: 45; readonly cachedInput: 1; }; }; readonly benchmarks: { readonly intelligence: 60; readonly gpqa: 93.5; }; readonly performance: { readonly outputTokensPerSecond: 63; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-5.6-luna': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.6 Luna"; readonly subtype: "reasoning"; readonly description: "OpenAI's smallest GPT-5.6 model. Efficient reasoning at the lowest cost with 1M context window and 128K max output. Excellent value for reasoning tasks. MMLU-Pro 78.4%, GPQA 92.3%, SWE-bench 62.7%. Supports tool use and sampling control."; readonly pricing: { readonly input: 0.2; readonly output: 1.2; readonly cachedInput: 0.02; }; readonly benchmarks: { readonly coding: 62.7; readonly mmluPro: 78.4; readonly gpqa: 92.3; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-5.6-sol': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.6 Sol"; readonly subtype: "reasoning"; readonly description: "OpenAI's largest GPT-5.6 model. Frontier-class reasoning with 1M context window and 128K max output. State-of-the-art performance on complex reasoning, coding, and agentic tasks. MMLU-Pro 83%, GPQA 94.6%, SWE-bench 64.6%. Supports tool use and sampling control."; readonly pricing: { readonly input: 4; readonly output: 20; readonly cachedInput: 0.4; }; readonly benchmarks: { readonly coding: 64.6; readonly mmluPro: 83; readonly gpqa: 94.6; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-5.6-terra': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-5.6 Terra"; readonly subtype: "reasoning"; readonly description: "OpenAI's mid-size GPT-5.6 model. Strong reasoning at lower cost with 1M context window and 128K max output. Competitive with GPT-5.5 on most benchmarks. MMLU-Pro 80.7%, GPQA 92.9%, SWE-bench 63.4%. Supports tool use and sampling control."; readonly pricing: { readonly input: 2; readonly output: 12; readonly cachedInput: 0.2; }; readonly benchmarks: { readonly coding: 63.4; readonly mmluPro: 80.7; readonly gpqa: 92.9; }; readonly contextLength: 1048576; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly topK: false; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-6-astra': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-6 Astra"; readonly description: "OpenAI's most capable model, built for the hardest end-to-end work. Achieves state-of-the-art performance on complex reasoning and agentic tasks. OSWorld 2.0 score 72.6%. Excels at autonomous task execution, complex reasoning, and knowledge-intensive applications. 1M context window with vision support."; readonly pricing: { readonly input: 10; readonly output: 50; readonly cachedInput: 1; }; readonly benchmarks: { readonly intelligence: 96; readonly coding: 74.1; readonly math: 97.6; readonly gpqa: 96; }; readonly contextLength: 1050000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-6-luna': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-6 Luna"; readonly description: "OpenAI's most efficient model for focused, high-volume tasks. Optimized for cost-effective inference with strong reasoning capabilities. Coding Agent Index score 41. Ideal for high-volume applications requiring fast, cost-effective reasoning. 1M context window with vision support."; readonly pricing: { readonly input: 0.1; readonly output: 0.5; readonly cachedInput: 0.01; }; readonly benchmarks: { readonly coding: 41; }; readonly contextLength: 1050000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-6-sol': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "GPT-6 Sol"; readonly description: "Built to power complex coding and agentic workflows. Delivers near-Astra performance at significantly lower cost. OSWorld 2.0 score 60.5%, Coding Agent Index 60. Strong performance on autonomous task execution and complex reasoning. 1M context window with vision support."; readonly pricing: { readonly input: 2; readonly output: 10; readonly cachedInput: 0.2; }; readonly benchmarks: { readonly coding: 60; }; readonly contextLength: 1050000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: false; }; }; readonly sampling: { readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-audio': { readonly type: "audio"; readonly provider: "openai"; readonly displayName: "GPT Audio"; readonly description: "GPT Audio model for audio understanding and processing. Supports text and audio input/output. Ideal for audio processing and generation applications."; readonly pricing: { readonly input: 2.5; readonly output: 10; readonly cachedInput: 0.25; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: true; }; readonly output: { readonly text: true; readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-23"; }; readonly 'gpt-audio-1.5': { readonly type: "audio"; readonly provider: "openai"; readonly displayName: "GPT Audio 1.5"; readonly description: "GPT Audio 1.5 model for audio understanding. Supports text and audio input/output. Ideal for audio processing and generation tasks."; readonly pricing: { readonly input: 4; readonly output: 16; readonly cachedInput: 0.4; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: true; }; readonly output: { readonly text: true; readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-23"; }; readonly 'gpt-audio-2025-08-28': { readonly type: "audio"; readonly provider: "openai"; readonly displayName: "GPT Audio (2025-08-28)"; readonly description: "GPT Audio August 2025 snapshot. Audio understanding and processing model. Ideal for audio applications."; readonly pricing: { readonly input: 2.5; readonly output: 10; readonly cachedInput: 0.25; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: true; }; readonly output: { readonly text: true; readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-21"; }; readonly 'gpt-audio-mini': { readonly type: "audio"; readonly provider: "openai"; readonly displayName: "GPT Audio Mini"; readonly description: "Compact GPT Audio model. Supports text and audio input/output. Ideal for cost-effective audio processing applications."; readonly pricing: { readonly input: 0.6; readonly output: 2.4; readonly cachedInput: 0.06; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: true; }; readonly output: { readonly text: true; readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-21"; }; readonly 'gpt-audio-mini-2025-10-06': { readonly type: "audio"; readonly provider: "openai"; readonly displayName: "GPT Audio Mini (2025-10-06)"; readonly description: "GPT Audio Mini October 2025 snapshot. Compact audio model. Ideal for cost-effective audio processing."; readonly pricing: { readonly input: 0.6; readonly output: 2.4; readonly cachedInput: 0.06; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: true; }; readonly output: { readonly text: true; readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-21"; }; readonly 'gpt-audio-mini-2025-12-15': { readonly type: "audio"; readonly provider: "openai"; readonly displayName: "GPT Audio Mini (2025-12-15)"; readonly description: "GPT Audio Mini December 2025 snapshot. Compact audio model for cost-effective audio processing."; readonly pricing: { readonly input: 0.6; readonly output: 2.4; readonly cachedInput: 0.06; }; readonly contextLength: 128000; readonly context: { readonly input: { readonly text: true; readonly audio: true; }; readonly output: { readonly text: true; readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-23"; }; readonly 'gpt-image-1': { readonly type: "image"; readonly provider: "openai"; readonly displayName: "GPT Image 1"; readonly description: "OpenAI's image generation model. Supports image input tokens ($2.5/1M) and output tokens ($8/1M). Ideal for image generation and editing tasks."; readonly unitPricing: { readonly per_1m_image_input_tokens: 2.5; readonly per_1m_image_output_tokens: 8; }; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly image: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-08-16"; }; readonly 'gpt-image-1-mini': { readonly type: "image"; readonly provider: "openai"; readonly displayName: "GPT Image 1 Mini"; readonly description: "Compact image generation model. Supports image input tokens ($2.5/1M) and output tokens ($8/1M). Ideal for cost-effective image generation."; readonly unitPricing: { readonly per_1m_image_input_tokens: 2.5; readonly per_1m_image_output_tokens: 8; }; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly image: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'gpt-image-1.5': { readonly type: "image"; readonly provider: "openai"; readonly displayName: "GPT Image 1.5"; readonly description: "State-of-the-art image generation model. Supports both text input tokens ($5/1M) and image input tokens ($8/1M) for image editing and manipulation. Ideal for advanced image generation and editing workflows."; readonly pricing: { readonly input: 5; readonly output: 10; readonly cachedInput: 0.5; }; readonly unitPricing: { readonly per_1m_input_tokens: 5; readonly per_1m_output_tokens: 10; readonly per_1m_image_input_tokens: 8; readonly per_1m_image_output_tokens: 32; }; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly image: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'gpt-image-2': { readonly type: "image"; readonly provider: "openai"; readonly displayName: "GPT Image 2"; readonly description: "OpenAI's state-of-the-art image generation model with 2K resolution and multilingual text rendering. Advanced image understanding and generation capabilities. Ideal for high-quality image generation and editing tasks."; readonly pricing: { readonly input: 5; readonly output: 30; readonly cachedInput: 0.5; }; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly output: { readonly text: false; readonly audio: false; readonly image: true; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-18"; }; readonly 'gpt-image-2-2026-04-21': { readonly type: "image"; readonly provider: "openai"; readonly displayName: "GPT Image 2 (2026-04-21)"; readonly description: "Dated snapshot of GPT Image 2, OpenAI's advanced image generation model. 2K resolution with multilingual text rendering. Ideal for professional image generation and editing."; readonly pricing: { readonly input: 5; readonly output: 30; readonly cachedInput: 0.5; }; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly output: { readonly text: false; readonly audio: false; readonly image: true; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-27"; }; readonly 'gpt-image-2.5-flare': { readonly type: "image"; readonly provider: "openai"; readonly displayName: "GPT Image 2.5 Flare"; readonly description: "OpenAI's image generation model with token-based pricing"; readonly pricing: { readonly input: 8; readonly output: 30; readonly cachedInput: 2; }; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-13"; }; readonly 'gpt-image-2.5-flare-2026-09-08': { readonly type: "image"; readonly provider: "openai"; readonly displayName: "GPT Image 2.5 Flare (2026-09-08)"; readonly description: "Dated snapshot of GPT Image 2.5 Flare"; readonly pricing: { readonly input: 8; readonly output: 30; readonly cachedInput: 2; }; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-13"; }; readonly 'gpt-image-2.5-sunburst': { readonly type: "image"; readonly provider: "openai"; readonly displayName: "GPT Image 2.5 Sunburst"; readonly description: "OpenAI's image generation model with token-based pricing"; readonly pricing: { readonly input: 8; readonly output: 30; readonly cachedInput: 2; }; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-13"; }; readonly 'gpt-image-2.5-sunburst-2026-09-08': { readonly type: "image"; readonly provider: "openai"; readonly displayName: "GPT Image 2.5 Sunburst (2026-09-08)"; readonly description: "Dated snapshot of GPT Image 2.5 Sunburst"; readonly pricing: { readonly input: 8; readonly output: 30; readonly cachedInput: 2; }; readonly context: {}; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-13"; }; readonly 'gpt-live-transcribe': { readonly type: "stt"; readonly provider: "openai"; readonly displayName: "GPT Live Transcribe"; readonly description: "OpenAI's live/streaming speech-to-text transcription model."; readonly unitPricing: { readonly per_minute: 0.017; }; readonly context: { readonly input: { readonly audio: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-realtime-translate': { readonly type: "audio"; readonly provider: "openai"; readonly displayName: "GPT Realtime Translate"; readonly subtype: "realtime"; readonly description: "Live translation model for real-time speech translation. Supports audio input with text and audio output. Pricing: $0.034/min. Ideal for real-time multilingual communication."; readonly unitPricing: { readonly per_minute: 0.034; }; readonly context: { readonly input: { readonly text: false; readonly audio: true; readonly image: false; readonly video: false; }; readonly output: { readonly text: true; readonly audio: true; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-realtime-whisper': { readonly type: "stt"; readonly provider: "openai"; readonly displayName: "GPT Realtime Whisper"; readonly description: "Streaming speech-to-text model for real-time transcription. Pricing: $0.017/min. Ideal for live transcription and real-time speech processing applications."; readonly unitPricing: { readonly per_minute: 0.017; }; readonly context: { readonly input: { readonly text: false; readonly audio: true; readonly image: false; readonly video: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'gpt-transcribe': { readonly type: "stt"; readonly provider: "openai"; readonly displayName: "GPT Transcribe"; readonly description: "OpenAI's next-generation speech-to-text transcription model."; readonly unitPricing: { readonly per_minute: 0.0045; }; readonly context: { readonly input: { readonly audio: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly o1: { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "o1"; readonly subtype: "reasoning"; readonly description: "OpenAI's reasoning-focused model with test-time compute. State-of-the-art on challenging benchmarks: GPQA Diamond 75.7%, MMLU-Pro 91.8%, math 74.3% (AIME pass rate). Codeforces ELO 1673 (89th percentile). 200K context window. Ideal for complex reasoning, mathematics, and competitive programming."; readonly pricing: { readonly input: 15; readonly output: 60; readonly cachedInput: 7.5; }; readonly benchmarks: { readonly coding: 89; readonly math: 94.8; readonly mmluPro: 90.8; readonly gpqa: 77.3; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: false; readonly temperature: false; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly sampling: { readonly topK: false; readonly topP: false; readonly temperature: false; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: false; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'o1-2024-12-17': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "o1 (2024-12-17)"; readonly subtype: "reasoning"; readonly description: "o1 December 2024 snapshot with test-time compute reasoning. GPQA 77.3%, MMLU-Pro 90.8%, math 94.8%. 200K context window. Coding 89%. Ideal for complex reasoning and mathematics."; readonly pricing: { readonly input: 15; readonly output: 60; readonly cachedInput: 7.5; }; readonly benchmarks: { readonly coding: 89; readonly math: 94.8; readonly mmluPro: 90.8; readonly gpqa: 77.3; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly o3: { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "o3"; readonly subtype: "reasoning"; readonly description: "OpenAI's latest reasoning model with state-of-the-art performance. AIME 2025: 98.4% with Python, GPQA 82.8%, MMLU-Pro 92.9%. Codeforces state-of-the-art. 200K context window. Optimized for complex reasoning, mathematics, and competitive programming. Always-on reasoning with configurable effort levels."; readonly pricing: { readonly input: 2; readonly output: 8; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly intelligence: 38; readonly coding: 81.3; readonly math: 98.4; readonly mmluPro: 92.9; readonly gpqa: 82.8; }; readonly performance: { readonly outputTokensPerSecond: 98; readonly timeToFirstToken: 7.89; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: true; readonly video: false; }; readonly tools: { readonly supported: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: false; readonly temperature: false; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly sampling: { readonly topK: false; readonly topP: false; readonly temperature: false; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: false; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'o3-2025-04-16': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "o3 (2025-04-16)"; readonly subtype: "reasoning"; readonly description: "o3 April 2025 snapshot with advanced reasoning capabilities. GPQA 87.7%, MMLU-Pro 91.6%, math 96.7%, coding 81.3%. 200K context window."; readonly pricing: { readonly input: 2; readonly output: 8; readonly cachedInput: 0.5; }; readonly benchmarks: { readonly coding: 81.3; readonly math: 96.7; readonly mmluPro: 91.6; readonly gpqa: 87.7; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'o3-mini': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "o3-mini"; readonly subtype: "reasoning"; readonly description: "OpenAI's efficient reasoning model optimized for STEM tasks and complex problem-solving. Matches o1's performance on challenging reasoning benchmarks like AIME (87.3%) and GPQA (77.2%) with medium reasoning effort, while running 24% faster. Ideal for cost-conscious applications requiring advanced reasoning. 200K context window. Always-on reasoning."; readonly pricing: { readonly input: 1.1; readonly output: 4.4; readonly cachedInput: 0.55; }; readonly benchmarks: { readonly coding: 66.7; readonly math: 87.3; readonly mmluPro: 86.9; readonly gpqa: 77.2; }; readonly performance: { readonly outputTokensPerSecond: 231.6; readonly timeToFirstToken: 4.76; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly tools: { readonly supported: false; }; readonly output: { readonly text: true; readonly audio: false; readonly image: false; readonly video: false; }; readonly sampling: { readonly topK: false; readonly topP: false; readonly temperature: false; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly sampling: { readonly topK: false; readonly topP: false; readonly temperature: false; readonly presencePenalty: false; readonly frequencyPenalty: false; }; readonly tools: { readonly supported: false; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'o3-mini-2025-01-31': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "o3 Mini (2025-01-31)"; readonly subtype: "reasoning"; readonly description: "o3 Mini January 2025 snapshot with efficient reasoning. GPQA 79.7%, MMLU-Pro 79.1%, math 87.3%. 200K context window. Optimized for cost-effective reasoning on STEM tasks."; readonly pricing: { readonly input: 1.1; readonly output: 4.4; readonly cachedInput: 0.55; }; readonly benchmarks: { readonly coding: 61; readonly math: 87.3; readonly mmluPro: 86.9; readonly gpqa: 79.7; }; readonly performance: { readonly outputTokensPerSecond: 231.6; readonly timeToFirstToken: 4.76; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'o4-mini': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "o4 Mini"; readonly subtype: "reasoning"; readonly description: "OpenAI's efficient reasoning model. Near o3-level performance at o3-mini pricing. AIME 98.9%, GPQA 78.4%, MMLU-Pro 83.2%. 200K context with vision."; readonly pricing: { readonly input: 1.1; readonly output: 4.4; readonly cachedInput: 0.275; }; readonly benchmarks: { readonly coding: 72.8; readonly math: 98.9; readonly mmluPro: 90; readonly gpqa: 78.4; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-20"; }; readonly 'o4-mini-2025-04-16': { readonly type: "chat"; readonly provider: "openai"; readonly displayName: "o4 Mini (2025-04-16)"; readonly subtype: "reasoning"; readonly description: "o4 Mini April 2025 snapshot with state-of-the-art reasoning efficiency. GPQA 78.4%, MMLU-Pro 83.2%, math 98.9%. 200K context with vision."; readonly pricing: { readonly input: 1.1; readonly output: 4.4; readonly cachedInput: 0.275; }; readonly benchmarks: { readonly coding: 72.8; readonly math: 98.9; readonly mmluPro: 90; readonly gpqa: 78.4; }; readonly contextLength: 200000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-23"; }; readonly 'omni-moderation-2024-09-26': { readonly type: "moderation"; readonly provider: "openai"; readonly displayName: "Omni Moderation (2024-09-26)"; readonly description: "OpenAI's multimodal content moderation model. Free to use."; readonly pricing: { readonly input: 0; readonly output: 0; readonly cachedInput: 0; }; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-25"; }; readonly 'omni-moderation-latest': { readonly type: "moderation"; readonly provider: "openai"; readonly displayName: "Omni Moderation Latest"; readonly description: "Alias for the latest OpenAI multimodal moderation model. Free to use."; readonly pricing: { readonly input: 0; readonly output: 0; readonly cachedInput: 0; }; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'sora-2': { readonly type: "video"; readonly provider: "openai"; readonly displayName: "Sora 2"; readonly description: "OpenAI's advanced video generation model. Creates high-quality videos from text prompts with improved motion, consistency, and visual fidelity. Supports up to 1080p resolution. Pricing: $0.10 per second of video generated at 720p. Ideal for content creation and video production."; readonly unitPricing: { readonly per_second: { readonly '720p': 0.1; }; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly video: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-24"; }; readonly 'sora-2-pro': { readonly type: "video"; readonly provider: "openai"; readonly displayName: "Sora 2 Pro"; readonly description: "Professional video generation model with higher resolution support. Supports 720p ($0.30/sec), 1024p ($0.50/sec), and 1080p ($0.70/sec) resolutions. Advanced features for professional video production and content creation."; readonly unitPricing: { readonly per_second: { readonly '720p': 0.3; readonly '1024p': 0.5; readonly '1080p': 0.7; }; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly video: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-24"; }; readonly 'text-embedding-3-large': { readonly type: "embeddings"; readonly provider: "openai"; readonly displayName: "Text Embedding 3 Large"; readonly description: "OpenAI's state-of-the-art text embedding model. MTEB score 64.6, placing it among the top embedding models. Supports 8,191 token context. Ideal for semantic search, RAG, and similarity tasks requiring high-quality embeddings."; readonly pricing: { readonly input: 0.13; readonly output: 0; readonly cachedInput: 0.013000000000000001; }; readonly benchmarks: { readonly intelligence: 64.6; }; readonly contextLength: 8191; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'text-embedding-3-small': { readonly type: "embeddings"; readonly provider: "openai"; readonly displayName: "Text Embedding 3 Small"; readonly description: "OpenAI's efficient text embedding model. Smaller and faster than Text Embedding 3 Large while maintaining strong performance. 8,191 token context. Ideal for cost-sensitive embedding applications."; readonly pricing: { readonly input: 0.02; readonly output: 0; readonly cachedInput: 0.002; }; readonly contextLength: 8191; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'text-embedding-ada-002': { readonly type: "embeddings"; readonly provider: "openai"; readonly displayName: "Text Embedding Ada 002"; readonly description: "Legacy text embedding model"; readonly pricing: { readonly input: 0.1; readonly output: 0; readonly cachedInput: 0.010000000000000002; }; readonly contextLength: 8191; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'tts-1': { readonly type: "tts"; readonly provider: "openai"; readonly displayName: "TTS-1"; readonly description: "Text-to-speech model optimized for speed"; readonly unitPricing: { readonly per_1m_characters: 15; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-07-28"; }; readonly 'tts-1-1106': { readonly type: "tts"; readonly provider: "openai"; readonly displayName: "TTS-1 (1106)"; readonly description: "TTS-1 November 2023 snapshot"; readonly unitPricing: { readonly per_1m_characters: 15; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-25"; }; readonly 'tts-1-hd': { readonly type: "tts"; readonly provider: "openai"; readonly displayName: "TTS-1 HD"; readonly description: "High-definition text-to-speech model"; readonly unitPricing: { readonly per_1m_characters: 30; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-08-26"; }; readonly 'tts-1-hd-1106': { readonly type: "tts"; readonly provider: "openai"; readonly displayName: "TTS-1 HD (1106)"; readonly description: "TTS-1 HD November 2023 snapshot"; readonly unitPricing: { readonly per_1m_characters: 30; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly audio: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-09-23"; }; readonly 'whisper-1': { readonly type: "stt"; readonly provider: "openai"; readonly displayName: "Whisper 1"; readonly description: "Whisper speech-to-text model"; readonly unitPricing: { readonly per_minute: 0.006; }; readonly context: { readonly input: { readonly audio: true; }; readonly output: { readonly text: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-07-28"; }; readonly 'grok-4.20-0309-non-reasoning': { readonly type: "chat"; readonly provider: "xai"; readonly displayName: "Grok 4.20 Non-Reasoning"; readonly description: "Tool-focused variant of Grok 4.20 optimized for speed and low-latency responses. Skips reasoning steps to deliver industry-leading speed for automated pipelines, coding agents, and real-time applications while retaining 1M token context capabilities."; readonly pricing: { readonly input: 1.25; readonly output: 2.5; readonly cachedInput: 0.2; readonly longContext: { readonly threshold: 200000; readonly input: 2.5; readonly output: 5; readonly cachedInput: 0.4; }; }; readonly benchmarks: { readonly intelligence: 22; }; readonly performance: { readonly outputTokensPerSecond: 163.8; readonly timeToFirstToken: 0.55; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly reasoning: {}; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'grok-4.20-0309-reasoning': { readonly type: "chat"; readonly provider: "xai"; readonly displayName: "Grok 4.20 Reasoning"; readonly subtype: "reasoning"; readonly description: "xAI's advanced reasoning variant with strong agentic tool calling and the lowest hallucination rate on the market. Features extended thinking process before answering for consistently precise and truthful responses. 1M token context window with long-context pricing tier at 200K tokens."; readonly pricing: { readonly input: 1.25; readonly output: 2.5; readonly cachedInput: 0.2; readonly longContext: { readonly threshold: 200000; readonly input: 2.5; readonly output: 5; readonly cachedInput: 0.4; }; }; readonly benchmarks: { readonly intelligence: 37; readonly coding: 78; readonly math: 96.5; readonly mmluPro: 91.2; readonly gpqa: 74.5; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly reasoning: {}; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'grok-4.20-multi-agent-0309': { readonly type: "chat"; readonly provider: "xai"; readonly displayName: "Grok 4.20 Multi-Agent"; readonly subtype: "reasoning"; readonly description: "Multi-agent variant of Grok 4.20 using a 4-agent 'Society of Mind' architecture for complex problem-solving. Combines multiple reasoning perspectives for enhanced analysis and decision-making. 1M token context window with long-context pricing tier."; readonly pricing: { readonly input: 1.25; readonly output: 2.5; readonly cachedInput: 0.2; readonly longContext: { readonly threshold: 200000; readonly input: 2.5; readonly output: 5; readonly cachedInput: 0.4; }; }; readonly benchmarks: { readonly coding: 44.7; readonly gpqa: 88.5; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly reasoning: {}; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'grok-4.3': { readonly type: "chat"; readonly provider: "xai"; readonly displayName: "Grok 4.3"; readonly subtype: "reasoning"; readonly description: "xAI's reasoning model balancing high reasoning density with production-ready efficiency. Delivers precise, multi-step responses for coding, research, and complex document workflows. Supports text and image inputs with configurable reasoning. 1M token context window."; readonly pricing: { readonly input: 1.25; readonly output: 2.5; readonly cachedInput: 0.2; readonly longContext: { readonly threshold: 200000; readonly input: 2.5; readonly output: 5; readonly cachedInput: 0.4; }; }; readonly benchmarks: { readonly intelligence: 38; readonly coding: 42.2; readonly gpqa: 90.1; }; readonly performance: { readonly outputTokensPerSecond: 134.5; readonly timeToFirstToken: 18.7; }; readonly contextLength: 1000000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly reasoning: {}; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'grok-4.5': { readonly type: "chat"; readonly provider: "xai"; readonly displayName: "Grok 4.5"; readonly subtype: "reasoning"; readonly description: "xAI's advanced reasoning model with strong performance on software engineering and coding tasks. Achieves 64.7% on SWE-bench Pro and 83.3% on Terminal Bench 2.1. 500K context window. Optimized for complex reasoning and code generation. Competitive pricing at $2/1M input tokens."; readonly pricing: { readonly input: 2; readonly output: 6; readonly cachedInput: 0.3; readonly longContext: { readonly threshold: 200000; readonly input: 4; readonly output: 12; readonly cachedInput: 0.6; }; }; readonly benchmarks: { readonly intelligence: 54; readonly coding: 57.3; readonly gpqa: 93.1; }; readonly performance: { readonly outputTokensPerSecond: 59.7; readonly timeToFirstToken: 10.38; }; readonly contextLength: 500000; readonly context: { readonly input: { readonly text: true; readonly image: true; readonly video: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'grok-4.6': { readonly type: "chat"; readonly provider: "xai"; readonly displayName: "Grok 4.6"; readonly subtype: "reasoning"; readonly description: "xAI's flagship model for code and everything else: agentic tool calling, minimal hallucinations, configurable reasoning. 500K token context window."; readonly pricing: { readonly input: 2; readonly output: 6; readonly cachedInput: 0.5; readonly longContext: { readonly threshold: 200000; readonly input: 4; readonly output: 12; }; }; readonly benchmarks: { readonly intelligence: 61; }; readonly performance: { readonly outputTokensPerSecond: 63.9; }; readonly contextLength: 500000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'grok-4.7': { readonly type: "chat"; readonly provider: "xai"; readonly displayName: "Grok 4.7"; readonly description: "xAI's latest Grok model with strong reasoning capabilities."; readonly pricing: { readonly input: 2; readonly output: 6; readonly cachedInput: 0.5; readonly longContext: { readonly threshold: 200000; readonly input: 4; readonly output: 12; }; }; readonly benchmarks: { readonly coding: 79.4; readonly math: 61.9; readonly mmluPro: 87; readonly gpqa: 88; }; readonly contextLength: 500000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly temperature: true; }; readonly reasoning: { readonly alwaysOn: true; }; }; readonly sampling: { readonly temperature: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'grok-build-0.1': { readonly type: "chat"; readonly provider: "xai"; readonly displayName: "Grok Build"; readonly subtype: "code"; readonly description: "xAI's coding-focused model optimized for code generation, debugging, and software development tasks. Default model for the Grok Build coding agent and Cursor integration. 256K token context window with long-context pricing tier at 200K tokens."; readonly pricing: { readonly input: 1; readonly output: 2; readonly cachedInput: 0.2; readonly longContext: { readonly threshold: 200000; readonly input: 2; readonly output: 4; readonly cachedInput: 0.4; }; }; readonly benchmarks: { readonly coding: 70.8; }; readonly contextLength: 256000; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly tools: { readonly supported: true; }; readonly output: { readonly text: true; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly reasoning: {}; }; readonly sampling: { readonly topK: true; readonly topP: true; readonly temperature: true; readonly presencePenalty: true; readonly frequencyPenalty: true; }; readonly tools: { readonly supported: true; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-24"; }; readonly 'grok-imagine-image': { readonly type: "image"; readonly provider: "xai"; readonly displayName: "Grok Imagine Image"; readonly description: "xAI's image generation model for creating images from text prompts. Supports configurable aspect ratio, resolution (1K, 2K), and output count (up to 10 images per request). Includes natural language image editing with support for up to 3 reference images."; readonly unitPricing: { readonly provider_reported: true; }; readonly contextLength: 8000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly image: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; readonly 'grok-imagine-image-2.0': { readonly type: "image"; readonly provider: "xai"; readonly displayName: "Grok Imagine Image 2.0"; readonly description: "xAI's advanced image generation and editing model. Ranked #4 on Artificial Analysis at 1,153.66 Elo. Excels at text rendering (5.0/5.0), prompt adherence (4.87/5.0), and composition (4.67/5.0). Ideal for high-quality image generation and editing tasks."; readonly unitPricing: { readonly per_image: 0.02; }; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly image: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-04"; }; readonly 'grok-imagine-image-quality': { readonly type: "image"; readonly provider: "xai"; readonly displayName: "Grok Imagine Image Quality"; readonly description: "Premium variant of Grok Imagine with enhanced image quality and generation capabilities. Supports text-to-image generation, natural language image editing, multi-image editing for style transfer and scene composition. Configurable aspect ratio and resolution (1K, 2K)."; readonly unitPricing: { readonly per_image: 0.05; }; readonly contextLength: 8000; readonly context: { readonly input: { readonly text: true; }; readonly output: { readonly image: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-12"; }; readonly 'grok-imagine-video': { readonly type: "video"; readonly provider: "xai"; readonly displayName: "Grok Imagine Video"; readonly description: "xAI's video generation model for creating videos from text prompts or image references. Supports configurable duration, aspect ratio, and resolution. Includes video editing, reference-guided generation, and video extension capabilities."; readonly unitPricing: { readonly per_second: { readonly '480p': 0.05; readonly '720p': 0.07; }; }; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly video: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-12"; }; readonly 'grok-imagine-video-1.5': { readonly type: "video"; readonly provider: "xai"; readonly displayName: "Grok Imagine Video 1.5"; readonly description: "Enhanced video generation model supporting creation from text or image references with preset voices. Configurable duration (up to 15 seconds), aspect ratio, and resolution (up to 1080p). Features video editing, reference-guided generation, and video extension with native video-audio generation."; readonly unitPricing: { readonly per_second: { readonly '480p': 0.08; readonly '720p': 0.14; readonly '1080p': 0.25; }; }; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly audio: true; readonly video: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-12"; }; readonly 'grok-imagine-video-1.5-lite': { readonly type: "video"; readonly provider: "xai"; readonly displayName: "Grok Imagine Video 1.5 Lite"; readonly description: "xAI's lightweight video generation model, distilled from Grok Imagine Video 1.5. Faster and lower-cost video generation from text or image references. Supports 480p, 720p, and 1080p resolutions."; readonly unitPricing: { readonly per_second: 0.02; }; readonly context: { readonly input: { readonly text: true; readonly image: true; }; readonly output: { readonly video: true; }; }; readonly deprecated: false; readonly isAvailable: true; readonly lastUpdated: "2026-10-11"; }; }; export type ModelId = keyof typeof MODEL_REGISTRY; export type SupportedModel = ModelId; export type Provider = SupportedProvider; export {}; //# sourceMappingURL=model-registry.d.ts.map