OpenAI will remove GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini from its API on Jan 14, 2027, with three months' notice, pointing developers to gpt-6-luna

Why it matters

【Key points】

OpenAI added an Oct 14 entry to its API deprecations page: the following snapshots will be removed from the API on Jan 14, 2027 — GPT-4o (gpt-4o-2024-08-06, gpt-4o-2024-11-20), GPT-4o mini (gpt-4o-mini-2024-07-18), GPT-4.1 (gpt-4.1-2025-04-14) and GPT-4.1 mini (gpt-4.1-mini-2025-04-14). All four point to gpt-6-luna, the low-cost model launched with GPT-6 Sol on Sept 22. Fine-tunes of these models shut down on the same date.

The page already lists gpt-4o-2024-05-13 and gpt-4.1-nano going on Oct 23, 2026 (announced Apr 22) and the gpt-4o-audio / gpt-4o-realtime families on Jan 20, 2027; chatgpt-4o-latest left on Feb 17, 2026. So GPT-4o and GPT-4.1 text models exit the OpenAI API by mid-January 2027. #keep4o users reacted again on X.

【Verifiable】

OpenAI API deprecations page: https://developers.openai.com/api/docs/deprecations

Its notice policy: at least 6 months for generally available models, 3 months for specialized variants, as little as 2 weeks for previews, unless safety or compliance concerns require faster.

Shutdown dates are also exposed via GET /v1/models (shutdown_date field).

Related: GPT-6 Sol and Luna (e/669)

【Our read】

1. OpenAI's own rule gives GA models six months; Oct 14 to Jan 14 is three, and the entry cites no safety or compliance reason. The April batch got the full six months. Until OpenAI explains, production teams should plan on three months.

2. GPT-4o mini and GPT-4.1 mini were default cheap models for many apps, with many fine-tunes on top. Those fine-tunes going too means re-tuning or re-prompting; the real cost is re-running evals, not editing a model string.

3. Sending all four to gpt-6-luna consolidates low-price traffic onto GPT-6, and closes the last official API route for users attached to 4o's style.