OpenAI Replaces ChatGPT's Default with GPT-5.5 Instant, Cutting Hallucinations by 52.5%

A New Default Lands in ChatGPT

OpenAI on Monday began rolling out GPT-5.5 Instant as the new default model powering ChatGPT, replacing GPT-5.3 Instant just two months after it became the standard on March 3rd. The release focuses on three main pillars: reducing factual errors, curbing excessive formatting, and deepening personalisation through a new transparency feature called 'memory sources'.

Fewer Hallucinations, Fewer Emojis

In internal evaluations, OpenAI says GPT-5.5 Instant produced 52.5% fewer hallucinated claims than its predecessor on high-stakes prompts covering medicine, law, and finance. The company also reported a 37.3% reduction in inaccurate claims on conversations that users had previously flagged for factual errors. OpenAI described the model as 'more dependable, with significant improvements in factuality across the board'.

Beyond accuracy, the update tackles a longstanding user complaint: cluttered, emoji-heavy outputs. OpenAI said GPT-5.5 Instant is designed to generate more direct responses and ask fewer unnecessary follow-up questions, while keeping the warmth and personality that makes ChatGPT enjoyable to use.

Personalisation and Memory Sources

The release also introduces deeper personalisation. GPT-5.5 Instant can search a user's past conversations, uploaded files, and connected Gmail accounts to deliver more tailored responses, meaning users should need to repeat themselves less often.

Alongside the model, OpenAI is launching a 'memory sources' feature across all ChatGPT models, designed to show users where the chatbot drew its context from. Users can delete outdated sources or correct inaccuracies, and OpenAI noted that memory sources will not be visible to others if a user shares a chat. The company acknowledged the feature 'won't display every bit of information ChatGPT might have used' but said it is working to make it more comprehensive.

Rollout Plans

GPT-5.5 Instant's personalisation capabilities are arriving first for Plus and Pro subscribers on the web, with mobile access coming soon and Free, Go, Business, and Enterprise tiers set to follow in the coming weeks. The model itself is being deployed to all ChatGPT users from Monday, while paid users will retain access to GPT-5.3 Instant for three months. The release comes less than two weeks after OpenAI launched the full GPT-5.5 model on 23rd April.