OpenAI is rolling out an update to ChatGPT's default model, dubbed GPT-5.5 Instant, promising smarter, clearer, and more personalized interactions for its hundreds of millions of users. This upgrade aims to make everyday use of the AI assistant more efficient and enjoyable.
Smarter, More Accurate Responses
The new model boasts significant gains in factuality, with internal evaluations showing a 52.5% reduction in hallucinated claims on high-stakes prompts covering medicine, law, and finance compared to its predecessor, GPT-5.3 Instant. It also improved accuracy on challenging user-flagged conversations by 37.3%.
GPT-5.5 Instant demonstrates enhanced capabilities across everyday tasks, including better analysis of image uploads, improved STEM question answering, and more judicious use of web search to provide relevant information.
Personalization and User Control
Personalization is a key focus, with GPT-5.5 Instant now more adept at leveraging context from past chats, connected files, and services like Gmail. This allows for more relevant suggestions and tailored plans without requiring users to repeat information.
A significant new feature is "memory sources," which provide transparency into what context was used to personalize a response. Users can view, delete, or correct this information, offering greater control over their AI interactions.
Temporary chats are also available, which do not use or update memory, providing a privacy-focused option.
Availability
GPT-5.5 Instant is available now to all ChatGPT users as the default model and via the API as `chat-latest`. Paid users will retain access to GPT-5.3 Instant for three months. Enhanced personalization features are rolling out to Plus and Pro subscribers, with plans to expand to other tiers and platforms in the coming weeks.
