OpenAI is updating ChatGPT's default model to be smarter and more accurate, delivering clearer, more concise answers that feel better tailored to each user.
Because Instant serves as the daily driver for hundreds of millions of people, even small improvements have a significant impact. This update makes everyday interactions more useful and enjoyable: stronger and tighter answers across subject areas, a more natural conversational tone, and better use of previously shared context when personalization can help.
Smarter, More Accurate Answers with Less to Sort Through
Instant is now more dependable, with significant improvements in factuality across the board and the largest gains in domains where accuracy matters most. In OpenAI's internal evaluations, GPT-5.5 Instant produced 52.5% fewer hallucinated claims than GPT-5.3 Instant on high-stakes prompts covering areas like medicine, law, and finance. It also reduced inaccurate claims by 37.3% on especially challenging conversations that users had flagged for factual errors.
GPT-5.5 Instant is a generally smarter model that is more capable across everyday tasks, including improvements in analyzing photo and image uploads, answering STEM-related questions, and deciding when to use web search to provide a more useful answer.
For example, when presented with a handwritten algebra solution, GPT-5.5 Instant demonstrates stronger reasoning than its predecessor. Both models initially endorse the user's work, but upon verification GPT-5.5 Instant recovers from its initial mistake: it catches that the proposed solution fails when plugged back into the original equation, identifies the actual algebra error (incorrect term rearrangement), and uses the quadratic formula to arrive at the correct answer. GPT-5.3 Instant also catches the failed substitution but stops too early, incorrectly concluding there is no real solution rather than revisiting the algebra.
These improvements are reflected in gains on evaluations across visual reasoning, math, and science.
With this update, responses are tighter and more to-the-point without losing substance, while retaining the warmth and personality that makes ChatGPT enjoyable to use. The model delivers the same information, often with more utility than previous versions, while reducing the verbosity and overformatting that can make responses too long. It also asks fewer unnecessary follow-up questions and avoids clutter like gratuitous emojis.
For instance, when a user asks for casual advice on handling a chatty coworker, GPT-5.5 Instant uses 30.2% fewer words and 29.2% fewer lines. Its response is stronger because it gets the tone right-informal, practical, and workplace-safe without overexplaining. It provides usable scripts for different situations and frames the issue around boundaries rather than the coworker's personality. GPT-5.3 Instant offers a more complete response with a "what not to do" section, but it feels slightly overbuilt for a casual advice prompt.
More Personalized Responses and Controls
Instant is now more effective at using context from past chats, files, and Gmail (if connected), so answers feel more personally relevant while keeping the user in control. It intelligently decides when a response can be improved with additional personalization and is faster at searching past conversations to find the right context, reducing the need for users to repeat themselves. This is especially helpful for getting tailored suggestions and plans, or picking up where a user left off on ongoing work.
For example, when a user asks for a tea shop recommendation, GPT-5.5 Instant references past chats and connected data to provide more refined, highly personalized suggestions based on the user's known preferences. GPT-5.3 Instant takes into account the user's location but provides more generic, higher-level suggestions.
OpenAI is also introducing memory sources across all ChatGPT models, giving users visibility into what context was used to personalize responses along with new controls. When a response is personalized, users can see what context was used-such as saved memories or past chats-and delete or correct it if something is outdated or no longer relevant.
Memory sources are not shown to others if a chat is shared. Users remain in control of their memory: they can delete chats they no longer want cited, delete or change items in saved memories via settings, or use temporary chats that don't use or update memory.
Memory sources are designed to make personalization easier to understand, though they may not show every factor that shaped an answer. For example, they may display some of the most relevant past chats rather than all that were searched and referenced. OpenAI plans to continue improving this view over time.
Availability
GPT-5.5 Instant is rolling out starting today to all ChatGPT users, replacing GPT-5.3 Instant as the default model, and is available in the API as chat-latest. For paid users, GPT-5.3 Instant will remain available for three months through model configuration settings before being retired.
Enhanced personalization from past chats, files, and connected Gmail is rolling out to Plus and Pro users on the web and coming soon to mobile, with plans to expand to Free, Go, Business, and Enterprise tiers in the coming weeks. Memory sources are rolling out across all ChatGPT consumer plans on the web and soon on mobile. Availability of specific personalization sources may vary by region.