ChatGPT’s New Default Model Is Half as Likely to Mislead You on Medical and Financial Questions
OpenAI says the update cuts hallucinations by 52.5% in high-stakes prompts and adds memory sources, tighter responses and API access.
- On Monday, OpenAI released GPT-5.5 Instant, a new foundation model replacing the previous version as ChatGPT's default, arriving two months after the last upgrade on March 3.
- OpenAI claims significant accuracy improvements, with internal testing showing 52.5% fewer hallucinated claims on high-stakes topics like law, finance, and medicine, plus 37.3% fewer inaccurate claims on challenging conversations.
- The model achieved 81.2 on the AIME 2025 math test versus 65.4 for its predecessor, and 76.0 on the MMMU-Pro benchmark versus 69.2, while improving web search and Gmail integration capabilities.
- Unlike Claude Opus 4.7 from Anthropic, Instant is "available to everyone," with OpenAI planning to extend access to Free and Business users in the coming weeks.
- For developers, the model is available via API as chat-latest, while paid users retain access to the older version for three months before the predecessor is sunsetted.
39 Articles
39 Articles
OpenAI rolls out ChatGPT 5.5 Instant as default model: What's new for users
OpenAI just rolled out ChatGPT 5.5 Instant as the new default model for everyone. So, if you were using GPT-5.3 Instant, that’s getting pushed aside. Paid users can hang onto the old model for a few months, but after that, it’s gone. What makes the thing more attractive is that GPT-5.5 Instant is not locked behind a paywall. Anyone can use it, and honestly, that’s a big deal for free users. Improved accuracy and fewer errors Now, the big upgrad…
GPT-5.5 Instant shows you what it remembered — just not all of it
OpenAI updated the default model for ChatGPT to its new GPT-5.5 Instant, along with a new memory capability that finally shows which context shaped responses — at least some of them. This limitation signals that models are starting to create a second, incomplete memory observability layer that could conflict with existing audit systems and agent logs. GPT-5.5 Instant replaces GPT-5.3 Instant as the default ChatGPT model and is a version of its n…
OpenAI rolls out GPT-5.5 Instant as default ChatGPT model, promises more accurate responses
OpenAI has announced that it’s replacing the default model in ChatGPT with GPT-5.5 Instant, which it said produces faster responses, shorter answers, and fewer incorrect claims. The model follows GPT-5.3 Instant, introduced in March, and is designed for everyday tasks such as writing, analysis, and general queries. OpenAI states in a blog post that “accuracy” is the model’s core selling point. “Instant is now more dependable, with significant im…
Coverage Details
Bias Distribution
- 54% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
