OpenAI Enhances GPT-5.5 Instant, Announces Retirement of Older Models

OpenAI has announced significant updates to its ChatGPT platform, primarily centered around an enhanced GPT-5.5 Instant model and the strategic retirement of several older models. The GPT-5.5 Instant update, which applies to both ChatGPT and its API, is designed to deliver a superior user experience by refining the model's response style and overall quality. Users can expect more natural and easier-to-read outputs, with conversations that flow better and are more appropriately paced for practical assistance tasks. This update specifically addresses previous feedback regarding overly verbose or bullet-point-heavy responses, aiming for a more concise and human-like interaction. A notable change accompanying this update is the discontinuation of the "canvas" feature within GPT-5.5 Instant and GPT-5.5 Thinking. However, writing and coding functionalities are now directly integrated into chat responses through dedicated writing and code blocks, offering a more streamlined workflow. Paid users will have a limited window to continue using canvas through legacy models until those models are sunset. In a broader move to optimize its model offerings, OpenAI is also retiring several older models that have seen limited usage within ChatGPT. OpenAI o3 is scheduled to be retired from ChatGPT on August 26, 2026, following a 90-day sunset period. Similarly, GPT-4.5 will be phased out on June 27, 2026, after a 30-day sunset period. These models were previously available exclusively to paid users via model settings. OpenAI emphasizes that these changes are confined to the ChatGPT platform and will not impact the API. This strategic consolidation allows OpenAI to focus resources on developing and supporting its most advanced and capable models, ensuring that users have access to the best available AI technology.