# Changelog v0.4.3 - 2026-05-31

### Added
- **[nvidia]** `stepfun-ai/step-3.7-flash` (Step 3.7 Flash, S+, 256k ctx) — new model appeared on NVIDIA NIM free tier on May 29, 2026
- **[qwen]** `qwen3.7-max` (Qwen3.7 Max, S+, 80.0% SWE, 1M ctx) — new flagship model confirmed on OpenRouter API (created 2026-05-20)
- **[openrouter]** `moonshotai/kimi-k2.6:free` (Kimi K2.6, S+, 262k ctx) — newly available with :free tier on OpenRouter

### Removed
- **[cloudflare]** `@cf/qwen/qwq-32b` (QwQ 32B) — upstream deprecation: Groq removed `qwen-qwq-32b` in July 2025, Cloudflare has mirrored this removal
- **[openrouter]** `minimax/minimax-m2.5:free` (MiniMax M2.5) — free tier removed from OpenRouter
- **[openrouter]** `deepseek/deepseek-v4-flash:free` (DeepSeek V4 Flash) — free tier removed from OpenRouter
- **[openrouter]** `baidu/cobuddy:free` (Baidu CoBuddy) — free tier removed from OpenRouter

### Fixed
- **[github-models]** `meta/llama-4-scout-17b-16e-instruct` ctx 10M → 1M — GitHub Models caps context at 1M, not 10M

### Changed
- Model count: 155 → 153 (net -2)
  - nvidia: 26 → 27 (+1)
  - cloudflare: 14 → 13 (-1)
  - openrouter: 26 → 24 (-3+1)
  - qwen: 11 → 12 (+1)
- All other providers verified current (no changes needed)

### Notes
- **gemini-2.5-pro/flash/flash-lite** still active until **October 16, 2026** — schedule removal for next audit cycle
- **Mistral deprecations** in July/August 2026: devstral-2512, magistral-medium-2509 — schedule removal for next audit cycle
- **SambaNova**: `gemma-3-12b-it` is now PREVIEW (not production) — still usable but may be removed at short notice