OpenAI has replaced ChatGPT's default model with GPT-5.5 Instant, rolling out globally to all ChatGPT users. The update replaces GPT-5.3 Instant, which launched two months ago, and brings three changes to the experience hundreds of millions of people use every day: fewer hallucinations, more concise responses, and deeper personalization.

This is not a premium model launch. GPT-5.5 Instant is the default, the version that opens when a user starts a new chat. Default model changes affect every user at scale, not only those who seek out new capabilities.

What Changed in GPT-5.5 Instant

GPT-5.5 Instant is not a new model tier. It replaces what was already running as ChatGPT's default. The three changes below affect every user who opens ChatGPT, whether they notice the switch or not.

Fewer hallucinations

In internal evaluations, GPT-5.5 Instant produced 52.5% fewer hallucinated claims than GPT-5.3 Instant on high-stakes prompts covering medicine, law, and finance. It also reduced inaccurate claims by 37.3% on conversations users had previously flagged for factual errors.

Benchmark scores from OpenAI's HealthBench test:

• HealthBench (general medical questions): improved from 49.6 to 51.4
• HealthBench Professional (clinical use): improved from 32.9 to 38.4

These are internal evaluations, not independent audits, and should be interpreted with that limitation in mind.

More concise responses

GPT-5.5 Instant uses 30.2% fewer words and 29.2% fewer lines than its predecessor. OpenAI called out the reduction of gratuitous emojis and excessive formatting as part of this change, a direct response to user complaints that ChatGPT responses felt cluttered. When GPT-5.3 Instant launched in March, OpenAI's stated goal was to make ChatGPT feel less cringe. This release focuses that intent on accuracy and brevity.

Deeper personalization

GPT-5.5 Instant can now draw on past conversations, uploaded files, and connected Gmail accounts to deliver more relevant responses. OpenAI is also rolling out Memory Sources alongside this update:

• Users can delete or correct any outdated memory source at any time
• Shared chats do not expose memory sources to recipients
• Temporary chats remain available for sessions where memory should not be used

Smart Switching and the Thinking Model

GPT-5.5 Instant introduces automatic routing. When a prompt requires deeper reasoning, Instant can switch to GPT-5.5 Thinking without the user doing anything. For manual selection, the model picker offers Instant for fast everyday responses and Thinking for complex analytical tasks.

The default model is no longer a fixed tier. It routes requests based on complexity before delivering an answer.

What This Means for Developers

For developers, GPT-5.5 Instant is now available via the API as chat - latest. GPT-5.3 Instant remains accessible to paid users for three months through model configuration settings before retirement. Any pipeline pointing at chat - latest will now run on GPT-5.5 Instant and should account for the shift in response style and factuality.

Personalization Rollout

GPT-5.5 Instant is rolling out as the default for all users now. Enhanced personalization covering past chats, uploaded files, and connected Gmail is reaching Plus and Pro users on web first, with mobile following shortly. Memory Sources are expanding across all consumer plans on web, with mobile support coming next.

Free, Go, Business, and Enterprise users will receive personalization features in the coming weeks. Paid subscribers retain access to GPT-5.3 Instant for three months before it is retired.

Read More

• OpenAI Drops GPT-5.5 — The Most Capable Model in the Room Is Live Now
• OpenAI Introduces Workspace Agents in ChatGPT — Build AI Automations Without Writing a Single Line of Code
• OpenAI and Microsoft End Exclusivity — What Their Amended Deal Means for the Future of AI
• OpenAI Launches ChatGPT Images 2.0 — The Image Model That Thinks Before It Draws
• Google Gemini Now Generates Files — PDF, Word, Excel and More, Directly From Your Chat