Every provider we add, every model that goes free (or stops being free) — logged automatically, no spin.
Groq now offers MiniMax M2.7 to Enterprise customers only (official model page, Oct 21 2026).
No longer in Kilo Gateway's free model list (live catalog, Oct 13 2026).
No longer in Kilo Gateway's free model list (live catalog, Oct 13 2026).
No longer on SiliconFlow's pricing page.
Model page no longer exists on Replicate (404).
Not a valid Model Studio model id (absent from the official model list).
Retired by Alibaba Cloud Model Studio on 2026-10-10 (official retirement notice).
Retired by Alibaba Cloud Model Studio on 2026-10-10 (official retirement notice).
Retired by Alibaba Cloud Model Studio on 2026-10-10 (official retirement notice).
The 'mistral' alias now points to a paid-only model on Pollinations.
No longer in Ollama's cloud model catalog.
Removed from build.nvidia.com and the integrate.api.nvidia.com model list.
No longer in the Nous Portal model catalog (live /v1/models, Oct 2026).
The :free variant was removed from OpenRouter (live /api/v1/models, Oct 13 2026).
The :free variant was removed from OpenRouter; only the paid qwen/qwen3.8-27b remains (live /api/v1/models, Oct 13 2026). Reported by a user.
Moved to llm7.io's paid 'pro' tier.
Deprecated by Groq (console.groq.com/docs/deprecations); not in the current model list.
Deprecated: now routed automatically to gemini-3.6-flash and no longer selectable (ai.google.dev deprecations).
Deprecated: now routed automatically to gemini-3.8-flash and no longer selectable (ai.google.dev deprecations).
No longer served serverless on Fireworks (on-demand deployment only).
No longer served serverless on Fireworks (on-demand deployment only).
Deprecated by Cloudflare on 2025-10-01 (Workers AI changelog).
Report a model that's no longer free directly from its provider page — it shows up here the moment we confirm it.
Browse the Directory