OpenAI Retires GPT-4o from ChatGPT
OpenAI has announced the retirement of its GPT-4o and GPT-4.1 models from the consumer-facing ChatGPT product. The company explained the rapid iteration cycle as the reason for the change. However, both models remain available for enterprise developers via the API.
- The retirement, effective February 13, 2026, also includes the GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini models from the consumer product, consolidating the offering around the newer GPT-5.2 model. - According to OpenAI, the decision was driven by user migration, with over 99.9% of daily usage having already shifted to GPT-5.2. - This is the second time GPT-4o has been removed; it was first deprecated with the launch of GPT-5 in August 2025, but was temporarily restored for subscribers following user complaints about losing a valued tool. - Enterprise and business customers have a slightly longer transition period, retaining access to GPT-4o within their Custom GPTs until April 3, 2026. - Originally launched in May 2024, GPT-4o—short for "omni"—was OpenAI's first truly multimodal model, capable of processing and generating text, images, and audio natively. - Its successor family, GPT-4.1, was released in April 2025 with a developer focus, offering a significantly larger 1 million token context window and superior performance on coding benchmarks like SWE-bench. - The retirement has generated backlash from a niche user base who built workflows around the model's specific behaviors and, in some cases, formed emotional attachments. - OpenAI states that feedback from the period when GPT-4o access was restored directly informed the development of GPT-5.2, leading to improved customization and personality controls in the new flagship model.
