Shared from twixb · theverge.com

OpenAI claims ChatGPT’s new default model hallucinates way less

theverge.com·May 5, 2026

OpenAI's new default model, GPT-5.5 Instant, claims to significantly reduce hallucinations, producing 52.5% fewer inaccurate claims compared to its predecessor, GPT-5.3. The model also enhances personalization and response quality while minimizing the use of unnecessary emojis, with a rollout for all users beginning soon.

OpenAI's new GPT-5.5 Instant model shows significant improvements in reducing hallucinations and inaccuracies, producing 52.5% fewer hallucinated claims compared to its predecessor. This advancement is particularly relevant for high-stakes domains like medicine, law, and finance. As someone tracking AI developments, consider how these enhancements could impact AI deployment in critical applications, where factual accuracy is paramount.

Powered by twixb

Want more content like this?

twixb tracks your favorite blogs and social media, filters by keywords, and delivers personalized key learnings — straight to your inbox.

More from AI & Machine Learning News

Recent stories curated alongside this one.