OpenAI releases GPT-5.5 Instant as default ChatGPT model with 52.5% fewer hallucinations

OpenAI deployed GPT-5.5 Instant as the default ChatGPT model on May 5, 2026, replacing GPT-5.3 Instant which launched on March 3. According to OpenAI, the update reduces hallucinated claims by 52.5% on high-stakes prompts covering medicine, law, and finance in internal evaluations.

Accuracy improvements

The company claims GPT-5.5 Instant reduced inaccurate claims by 37.3% on conversations users previously flagged for factual errors. OpenAI describes these as "significant improvements in factuality across the board" with the largest gains in domains where accuracy is critical.

Response format changes

GPT-5.5 Instant delivers "tighter and more to-the-point" responses while reducing verbosity and formatting issues from previous versions. OpenAI specifically notes the model "avoids things that can make responses feel cluttered, like gratuitous emojis" and asks fewer unnecessary follow-up questions.

The update focuses on conciseness without losing substance, according to OpenAI, while maintaining "the warmth and personality that makes ChatGPT enjoyable to use."

Availability

The rollout began May 5 for all ChatGPT users. Paid subscribers will retain access to GPT-5.3 Instant for three months. Pricing and context window specifications were not disclosed in the announcement.

This follows OpenAI's April release of GPT-5.5 Thinking and Pro models, which the company positioned as "a new class of intelligence for real work."

What this means

OpenAI's claimed 52.5% reduction in hallucinations on high-stakes prompts represents a significant accuracy improvement if verified by independent testing. The focus on reducing verbosity and emoji usage suggests OpenAI is responding to user feedback about ChatGPT's output style becoming overly formatted. The three-month grace period for paid users to access the previous model indicates OpenAI anticipates some users may prefer GPT-5.3 Instant's response style.