Retiring GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini in ChatGPT

Source: OpenAI Date: 2026-01-29 URL: https://openai.com/index/retiring-gpt-4o-and-older-models

Summary

OpenAI announcement from January 2026 deprecating GPT-4o, GPT-4.1, GPT-4.1 mini, and o4-mini from the ChatGPT product surface — not from the API, but from the consumer and enterprise ChatGPT interfaces. The retirement consolidates ChatGPT around the GPT-5.x family as the standard offering, removing the prior-generation model options that had been available to subscribers. API customers typically receive longer deprecation windows than product surfaces.

Implications

Model lifecycle management. This signal is one of the clearest markers of OpenAI accelerating its model lifecycle. GPT-4o was the flagship model for most of 2024 and the backbone of the Apple Intelligence integration. Retiring it from ChatGPT in January 2026 — roughly 18 months after launch — reflects confidence that GPT-5.x is broadly superior for the use cases ChatGPT serves. The accelerating release cadence (GPT-5.1 through 5.4 in six months) made the older model tier harder to justify.

User migration pressure. Users who had tuned their workflows to GPT-4o’s specific characteristics (formatting behavior, context window tradeoffs, pricing) now face involuntary migration. This is a recurring pattern in the OpenAI product: model retirements generate friction for power users who adapted to specific model behaviors.

API vs. product divergence. The distinction between “retired from ChatGPT” and “deprecated from the API” matters for developers. Enterprise API users typically have longer runways; consumer product users are on OpenAI’s timeline. This divergence reflects the different durability requirements between consumer UX and production software integrations.

Watch: Whether GPT-4o deprecation from the API follows on a similar 3-6 month lag, and what migration path API customers with fine-tuned GPT-4o models are offered.