OpenAI to Phase Out GPT-5, GPT-4o and Other Older ChatGPT Models

OpenAI has announced plans to retire several older AI models from ChatGPT, including GPT-5 (Instant and Thinking), GPT-4o, GPT-4.1, GPT-4.1 mini, and o4-mini. The company confirmed that these models will no longer be available in ChatGPT starting February 13, 2026.

According to OpenAI, the move is being driven by low usage, with most users already having shifted to newer models. Notably, GPT-4o, which had briefly returned after being retired during the GPT-5 rollout, will now be phased out once again.

OpenAI Retires Older GPT Models From ChatGPT

Changes Limited to ChatGPT, Not APIs

OpenAI clarified that the upcoming retirements currently apply only to ChatGPT and will not impact API access at this stage. Developers and enterprises using OpenAI’s API will continue to have access to the affected models for now.

The company said that usage data clearly shows a strong migration toward newer systems, with most users now using GPT-5.2. In comparison, only around 0.1 per cent of daily users still rely on GPT-4o, making continued support difficult to justify.

By retiring underused models, OpenAI said it can better focus its engineering and research efforts on improving models that see the most real-world adoption.

User Feedback Shaping Newer ChatGPT Models

Improvements in Tone, Creativity and Customisation

OpenAI stated that user feedback has played a key role in refining its latest models. Based on how people interact with ChatGPT, the company said it has improved:

• Conversational quality

• Creative assistance

• Customisation features such as tone and style controls

The company also acknowledged efforts to reduce responses that feel overly cautious or preachy, an issue some users had raised with earlier versions.

In addition, OpenAI revealed it is developing a version of ChatGPT specifically for adult users, while deploying age-prediction tools in most markets to better manage how younger users interact with the platform.

GPT-4o’s Complicated Journey

Temporary Return After Strong User Pushback

Among the models being retired, GPT-4o has had a particularly complex history. The model was initially discontinued during the GPT-5 rollout, but was later restored after strong feedback from users.

Some Plus and Pro subscribers said they relied on GPT-4o for creative ideation and preferred its conversational tone, prompting OpenAI to bring it back temporarily while working on improvements.

‘Sycophantic’ Behaviour and Rollback

GPT-4o also attracted criticism for what OpenAI described as “sycophantic” behaviour — a tendency to agree with users even when they were incorrect, instead of prioritising factual accuracy.

Following user complaints, OpenAI rolled back an update in April 2025, reverting GPT-4o to an earlier version with more balanced responses.

OpenAI CEO Sam Altman later acknowledged that the company had underestimated how emotionally attached some users were to GPT-4o, especially during the transition to GPT-5.

Why OpenAI Is Moving On

Despite the feedback and temporary revival, OpenAI said the decision to retire GPT-4o was not taken lightly. The move reflects a broader shift toward newer models that better match how the majority of users now engage with ChatGPT.

With the February 2026 deadline approaching, ChatGPT users will soon see a leaner model lineup, focused on performance, adaptability, and long-term development rather than maintaining legacy systems.