OpenAI To Retire GPT-4o And Other Legacy Models From ChatGPT On February 13

OpenAI has retired some legacy artificial intelligence models from ChatGPT, including the popular GPT-4o. It is the second time for GPT-4o after an unsuccessful sunset in August 2025, where user outcry led to its reinstatement. At the time, users complained that newer models were missing GPT-4o’s emotional nuance and writing. Now, OpenAI says the newer models offer comparable capabilities, and also the daily usage of GPT-4o has fallen.

GPT-4o Retirement Date Confirmed

On 29 January, OpenAI announced in a blog post that they are retiring GPT-4o (along with the models GPT‑4.1, GPT‑4.1 mini, and OpenAI o4-mini). OpenAI specified that the majority of the models below will be taken down from ChatGPT on February 13th:

- GPT-4o

- GPT-5 Instant

- GPT-5 Thinking

- GPT-4.1

- GPT-4.1 mini

- o4-mini

This is a breaking change just for ChatGPT. Access to the models via the OpenAI API will not change.

ALSO READ: Apple To Delay Standard iPhone 18 Launch, Prioritise Premium And Foldable Models For 2026

Why OpenAI Is Retiring GPT-4o

OpenAI did not include some idiomatic language from the world of competitive typing that would have diluted GPT-40’s bias toward text about keyboards. 2. Improvements include:

- Personality fine-tuning

- Creative ideation support

- Customisation tools

- Preset response styles

- Adjustable warmth and enthusiasm levels

According to OpenAI, only 0.1 per cent of users now choose GPT-4o daily.

User Backlash on Social Media

Some users asked about OpenAI’s usage stats. One X user, Claire (@Claire20250311), said usage of GPT-4o had fallen largely as a result of GPT-5. 2 was moved to the default, and the GPT-4o is pay-to-use. A different user, @Clo0oOoud, noted that the thousands of trending posts are evidence of more interest than OpenAI would lead users to believe.

ALSO READ: CERT-In Warns macOS And Chrome Users Of High-Risk Vulnerabilities, Urges Immediate Updates

Writing Quality Concerns

Earlier this week, OpenAI CEO Sam Altman hinted at the possibility of GPT-5. As feature development priorities shifted in favour of coding, reasoning, and math for GPT-2, we also offended many users by neglecting GPT-2o’s strengths for writing performance.