GPT-5.5

OpenAI's most capable model. AA Intelligence Index: 60 (#1 globally). Described as "a new class of intelligence" that understands intent faster and carries more of the work itself. 1M-token context window (API), 400K in Codex. Excels at writing/debugging code, web research, data analysis, document creation, and multi-tool agentic workflows.

Terminal-Bench 2.0: 82.7% (#1, leading Opus 4.7 by 13pp), FrontierMath Tiers 1-3: 51.7%, plus SOTA on OSWorld-Verified, GDPval, and CyberGym. Available in ChatGPT, Codex, and API with multiple reasoning effort levels (low/medium/high/xhigh). Proprietary.