OpenAI discontinues API access to popular GPT-4o model starting February 2026

OpenAI has recently announced that its chatgpt-4o-latest model will be retired from the developer platform in mid-February 2026. The access to this model will be terminated on February 16, 2026, allowing a three-month transition period for applications still utilizing GPT-4o.

It’s important to note that this timeline only applies to the API, and there is no set schedule for removing GPT-4o from ChatGPT, where it remains available for individual consumers and users across various paid subscription tiers. While internally, the model is considered a legacy system with lower API usage compared to newer models like GPT-5.1, OpenAI will provide developers with ample warning before removing any models.

GPT-4o, also known as “Omni,” was released around 1.5 years ago in May 2024, introducing OpenAI’s first unified multimodal architecture. It enabled near real-time conversational speech and delivered significant improvements in image understanding, multilingual support, document analysis, and expressive voice interaction. The model quickly became popular among millions of ChatGPT users, offering various capabilities and features.

The retirement of GPT-4o has sparked backlash from users who have developed strong emotional connections with the model. Some users even relied on GPT-4o for personal support and everyday tasks, leading to a movement to keep the model alive. However, OpenAI’s decision to transition to newer models like GPT-5.1 is part of its product strategy to offer more advanced features and capabilities.

Developers who are still using GPT-4o are encouraged to migrate to GPT-5.1, which offers larger context windows, advanced reasoning modes, and higher throughput options. The retirement of GPT-4o also coincides with a restructuring of OpenAI’s API pricing, with newer models like GPT-5.1 being more cost-effective compared to GPT-4o.

Overall, the retirement of GPT-4o signifies the evolution of OpenAI’s ecosystem towards more powerful and efficient models. While the transition may require some adjustments for developers, it reflects the company’s commitment to innovation and improvement in AI technology.