OpenAI has swapped out GPT-5.3 Instant for GPT-5.5 Instant as the default ChatGPT model, rolling the update out to all users starting today. The new model focuses on improving three things: accuracy, conciseness, and personalization.
Less hallucination, less fluff
OpenAI says the accuracy improvements are most notable in high-stakes subject areas. In internal testing, GPT-5.5 Instant produced 52.5% fewer hallucinated claims than its predecessor on prompts covering medicine, law, and finance, and 37.3% fewer inaccurate claims on conversations users had previously flagged for errors.
The new model even handles photo analysis, STEM questions, and web search decisions better than before. Responses are also shorter and less cluttered, with OpenAI claiming that the new model uses 30.2% fewer words and 29.2% fewer lines compared to GPT-5.3 Instant.
Users should also notice fewer unnecessary follow-up questions and formatting.
Better memory, more control
GPT-5.5 Instant is also better at pulling in relevant context from past conversations, uploaded files, and connected Gmail accounts to personalize responses without requiring users to repeat themselves.
Alongside this, OpenAI is introducing “memory sources” across all ChatGPT models. This new feature will show users exactly what context was used to shape a response. Users will be able to delete or correct any of it directly from settings.
The GPT-5.5 Instant is rolling out over the next two days to all ChatGPT users. GPT-5.3 Instant will remain available to paid users for three months via model configuration settings before being retired. Enhanced personalization is rolling out first to Plus and Pro users on the web, with Free, Go, Business, and Enterprise to follow in the coming weeks.
Memory sources are rolling out across all ChatGPT consumer plans on the web. OpenAI says personalization improvements and memory sources will also roll out on mobile soon.