In an effort to optimize resources and concentrate on more recent, sophisticated systems, OpenAI has declared that it will stop using a number of older ChatGPT models on February 13, 2026. GPT-4o, GPT-4.1, GPT-4.1 Mini, and OpenAI o4-Mini are among the models that will be decommissioned; they are all now categorized as legacy models.
The business claims that almost 99.9% of consumers have already switched to more recent models, such GPT-5.2, with little use left for earlier iterations. According to OpenAI, resources that may be used to enhance performance and user experience on existing platforms are diverted by maintaining underutilized models.
GPT-4o has already faced retirement. After the release of GPT-5 last year, it was momentarily taken down, but paid users’ input led to its reinstatement. According to OpenAI, more recent models have included features that users found appealing, such as improved conversational tone, increased originality, and increased personalization.
Users can keep working on older versions until the cutoff in February 2026. Existing conversations will then immediately switch to GPT-5.2 or later versions, including picture generation.
According to OpenAI, the change is intended to give consumers a more seamless, customized, and human-like AI experience in the future.