Introducing GPT-5.5

OpenAI has launched GPT-5.5, its most intelligent and user-friendly model to date, designed to revolutionize how work is done on computers. GPT-5.5 demonstrates enhanced understanding and capability in tasks such as coding, online research, data analysis, document creation, and software operation. It can handle complex, multi-part tasks autonomously, planning, utilizing tools, and verifying its work. The model shows significant improvements in agentic coding, computer use, knowledge work, and scientific research, areas that require sustained reasoning and action. Notably, GPT-5.5 achieves this increased intelligence without sacrificing speed, matching GPT-5.4's per-token latency while offering superior performance and token efficiency. OpenAI has implemented robust safeguards to mitigate misuse while ensuring access for beneficial applications, following extensive testing and feedback from early partners. GPT-5.5 is now available to various user tiers in ChatGPT and Codex, with API access forthcoming. Benchmarks indicate GPT-5.5 outperforms previous models and competitors across coding, professional tasks, computer use, tool utilization, academic research, and cybersecurity evaluations. The model's advancements are attributed to its improved reasoning, autonomy, and ability to handle complex, long-horizon tasks, making it a powerful tool for both individual productivity and scientific discovery.

(Source：OpenAI)