OpenAI has begun rolling out GPT-5.5 Instant as the new default model for ChatGPT, promising greater factual accuracy and improved personalization. The update reduces hallucinations and inaccurate claims while enhancing response clarity. A new memory sources feature gives users more control over personalized context.
OpenAI started deploying GPT-5.5 Instant to all ChatGPT users on May 5, 2026, replacing the previous default model. The company claims the new model produces 52.5 percent fewer hallucinated claims and 37.3 percent fewer inaccurate claims in internal tests, particularly on flagged conversations. It also delivers more direct responses with less cluttered outputs, including reduced use of gratuitous emojis and fewer unnecessary follow-up questions, while maintaining a warm personality, according to OpenAI. GPT-5.5 Instant improves personalization by faster searching of uploaded files, connected Gmail accounts, and prior conversations, reducing the need for users to repeat information. OpenAI is introducing memory sources alongside the model, allowing users to view and manage the context used for personalized responses, such as saved memories or past chats. Users can delete or update outdated information, with temporary chats excluded from memory updates. Memory sources are private and not shown when sharing chats, OpenAI explains: “You remain in control of what's in your memory: you can delete chats you no longer want to be cited, delete or change items in saved memories in settings, or use temporary chats that don't use or update your memory.” The company notes the feature currently does not show every bit of used information but plans to make it more comprehensive. Plus and Pro subscribers receive the personalization upgrades first, with Free, Go, Business, and Enterprise users to follow in coming weeks. All users will get memory sources on web and mobile soon. GPT-5.3 Instant remains available to paid users for three months before retirement.
