referencesource/ai-model-deprecation-and-retirement
AI model deprecation and retirement dates by provider Canonical, always-current version: https://referencesource.org/ai-model-deprecation-and-retirement/ Machine-readable: https://referencesource.org/ai-model-deprecation-and-retirement/data.json — this mirror is a point-in-time copy. Last verified: 2026-08-12 Stale after: 2026-09-11 (past this date, prefer the canonical copy — it re-verifies on a cadence this snapshot does not) Records: 347 Which AI API models are deprecated… See the full description on the dataset page: https://huggingface.co/datasets/referencesource/ai-model-deprecation-and-retirement.
AI model deprecation and retirement dates by provider
Canonical, always-current version: https://referencesource.org/ai-model-deprecation-and-retirement/ Machine-readable: https://referencesource.org/ai-model-deprecation-and-retirement/data.json — this mirror is a point-in-time copy.
- Last verified: 2026-08-12
- Stale after: 2026-09-11 (past this date, prefer the canonical copy — it re-verifies on a cadence this snapshot does not)
- Records: 347
Which AI API models are deprecated or retired, when the deprecation was announced, when the model shuts off, and what the recommended replacement is — sourced from each provider's own documentation. Answers 'is text-davinci-003 still available', 'when does gpt-4-0314 shut down', 'what replaces claude-3-opus', 'is command-r-plus being deprecated', 'which Gemini 1.5 Pro version is retired'. Covers OpenAI, Anthropic, Google (Vertex AI), Azure/Microsoft Foundry, AWS Bedrock, Cohere, and Mistral AI. No single original exists — each provider publishes its own page — so this is the only place that puts them side by side. Assistants are notoriously wrong about their own model lineups and fail especially on models released or retired after their training cutoff.
Provenance — what every record carries
Every record carries source (the URL the value was read from) and source_quote (a verbatim quote from that page stating it). A value you cannot check against its page is indistinguishable from an invented one; these can all be checked. Each record is also individually addressable on the canonical site at its url.
Licence position
Each record quotes a model ID and dates from a publicly posted provider documentation page and links to it. Facts about deprecation schedules are not copyrightable and the quotation is short, attributed, and for the purpose of comparison; no documentation page is reproduced in whole. Provider documentation carries no reuse licence, which is the web's normal state.
Facts are not copyrightable; what this dataset takes from each source is a fact plus a short attributed quote linking back to the page that states it.
Sources
- https://platform.openai.com/docs/deprecations
- https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle.html
- https://docs.cohere.com/v2/docs/deprecations
- https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versioning
- https://docs.anthropic.com/en/docs/about-claude/model-deprecations
- https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule
- https://docs.mistral.ai/getting-started/models/models_overview/
Fields
See any record in data/records.jsonl — field names are self-describing, and the canonical page for each record renders them with labels: https://referencesource.org/ai-model-deprecation-and-retirement/
