Model retirements
Re-test before it's retired.
When a model is retired you have to move, and the replacement won't answer exactly as the old one did. These are the dates that matter now, each from the developer's own page. Checked 8 Nov 2026.
Models
| Date | Retiring | Official replacement | |
|---|---|---|---|
| 30 Nov 2026 22 days left |
Claude Sonnet 4.5 Anthropic · claude-sonnet-4-5-20250929 |
claude-sonnet-5-5 |
How to re-test → |
| 11 Dec 2026 33 days left |
GPT-5, GPT-5 mini, nano and pro OpenAI · gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-pro |
gpt-5.6-sol, gpt-5.6-terra, gpt-5.6-luna |
How to re-test → |
| 14 Jan 2027 67 days left |
GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini OpenAI · gpt-4o-2024-08-06, gpt-4o-2024-11-20, gpt-4o-mini, gpt-4.1, gpt-4.1-mini |
gpt-6-luna |
How to re-test → |
Tools shutting down
Some of the places people kept their prompts and evals are closing too.
| Date | What | Source |
|---|---|---|
| 30 Nov 2026 22 days left |
OpenAI Evals platform Announced on 3 June 2026. Existing evals became read-only on 31 October; the dashboard and API shut on 30 November. OpenAI's migration guide is “Moving from OpenAI Evals to Promptfoo”. What to do with your evals. |
OpenAI: deprecations |
| 30 Nov 2026 22 days left |
OpenAI reusable prompts (v1/prompts) and Agent Builder Both shut on 30 November 2026, the same day as Evals. |
OpenAI: deprecations |
| 17 Aug 2026 Closed |
Anthropic legacy Workbench: saved prompts and evals Legacy Workbench access ended. Saved prompts, variables and evals aren't supported in the updated Workbench and Playground, and the experimental generate_prompt, improve_prompt and templatize_prompt APIs were retired the same day. Where to test prompts now. |
Claude release notes, 17 July 2026 |