Here lie the models
we used to call.
A running record of every LLM that appeared on, vanished from, or quietly cost more on OpenRouter — so the next outage isn’t a surprise.
What this is
OpenRouter is the unified gateway thousands of apps use to call 300+ LLMs from Anthropic, OpenAI,
Google, Meta, Mistral, xAI, and others. Its /api/v1/models catalog mutates constantly — models appear, get retired, change price.
The problem: many models disappear silently. The canonical receipt is BerriAI/litellm #20521 — 39 OpenRouter models vanished from the live API while still appearing in
pricing JSONs, with no deprecation_date to warn downstream apps. ModelGraveyard fetches the catalog every 6 hours,
diffs it, and publishes the diff as typed events + an Atom feed. Read more →
Catalog cross-reference
how this worksScheduled funerals
Models that have announced a retirement date.
- openai/gpt-5.2-chat 2026-08-10
Latest events
- added 2026-05-27 openai/gpt-3.5-turbo
- added 2026-05-27 openai/gpt-4
- added 2026-05-27 openai/gpt-4-0314
- added 2026-05-27 gryphe/mythomax-l2-13b
- added 2026-05-27 undi95/remm-slerp-l2-13b
- added 2026-05-27 mancer/weaver
- added 2026-05-27 openai/gpt-3.5-turbo-16k
- added 2026-05-27 openai/gpt-3.5-turbo-instruct
- added 2026-05-27 mistralai/mistral-7b-instruct-v0.1
- added 2026-05-27 openai/gpt-4-1106-preview
- added 2026-05-27 openrouter/auto
- added 2026-05-27 openai/gpt-4-turbo-preview
- added 2026-05-27 openai/gpt-3.5-turbo-0613
- added 2026-05-27 mistralai/mistral-large
- added 2026-05-27 anthropic/claude-3-haiku
- added 2026-05-27 openai/gpt-4-turbo
- added 2026-05-27 microsoft/wizardlm-2-8x22b
- added 2026-05-27 mistralai/mixtral-8x22b-instruct
- added 2026-05-27 meta-llama/llama-3-8b-instruct
- added 2026-05-27 meta-llama/llama-3-70b-instruct