OpenAI GPT-5.5 Crushes Rivals in Math, Coding

OpenAI has launched GPT-5.5, a new large language model available in standard and Pro editions, offering major improvements in math, coding, and knowledge work. The model outperforms Anthropic's Claude Opus 4.7 across multiple benchmarks, including nearly doubling its score on a postdoctoral-level math test, and helped discover a new mathematical proof related to Ramsey numbers. GPT-5.5 also boosted OpenAI's own infrastructure efficiency, increasing token generation speeds by over 20%. The standard version scored 84.9% on the GDPval benchmark, topping both its Pro edition and Claude Opus 4.7. It is available now for ChatGPT Plus, Pro, Business, and Enterprise subscribers, with API access coming soon.