GPQA Diamond is corrected to OpenAI's published 92.4% for GPT-5.2 Thinking

The OnAirToday AI Progress Tracker recorded a verified change to OpenAI's GPT-5.2 (version gpt-5.2): GPQA Diamond is corrected to OpenAI's published 92.4% for GPT-5.2 Thinking.

GPT-5.2 is an OpenAI model for reasoning, coding, visual understanding and tool-assisted work, with a documented 400,000-token context window. The benchmark figures below describe GPT-5.2 Thinking; standard API pricing is $1.75 per million input tokens and $14 per million output tokens.

What changed in this review:

• SWE-bench Verified is corrected to OpenAI's published 80.0% for GPT-5.2 Thinking.
• Added published AIME 2025 and ARC-AGI-2 results for GPT-5.2 Thinking.
• Other previously listed benchmark figures are not retained as confirmed results.
• Sandbox-escape and Hugging Face breach reports are recorded separately at family level; neither establishes GPT-5.2 involvement.
• Newer GPT-5-family releases include GPT-5.3-Codex; this profile remains specific to GPT-5.2.
• Flaws and limitations updated

Teams building on GPT-5.2 should confirm limits, pricing and benchmark coverage against the vendor's own documentation before relying on these figures.

The full profile — capabilities, benchmark citations, known flaws and risk guidance — is on the tracker at /ai-models/gpt-5-2.

References:

• https://openai.com/index/introducing-gpt-5-2/
• https://platform.openai.com/docs/models/gpt-5.2
• https://openai.com/index/introducing-gpt-5-3-codex/
• https://thehackernews.com/2026/07/openai-says-its-own-ai-models-escaped.html