OpenAI has announced GPT-5.5.
This new AI model aims to improve performance in coding, data analysis, research, and general computer tasks.
According to OpenAI, GPT-5.5 can write and debug code better, work with documents and spreadsheets, browse and use tools, and carry out longer tasks without frequent instructions.
It is also designed to manage unclear or mixed instructions and continue working until a task is completed.
The model shows improved results in several internal and external benchmarks, especially in software engineering, tool use, and scientific-style problem solving.
OpenAI also says GPT-5.5 uses fewer tokens for many tasks compared to earlier models, which improves efficiency.
The company highlighted progress in “agent-like” behavior, where the model can plan steps, use tools, and adjust its approach during a task.
ChatGPT-5.5 Availability
The rollout of GPT-5.5 is starting for ChatGPT Plus, Pro, Business, and Enterprise users, with broader availability in Codex.
A higher-performance version, GPT-5.5 Pro, is also being released for paid and enterprise users. API access is expected soon, with pricing to be announced separately.