AI Research

    GPQA Diamond now reports OpenAI's published 92.4% GPT-5.2 Thinking result

    GPQA Diamond now reports OpenAI's published 92.4% GPT-5.2 Thinking result — verified by the OnAirToday AI Progress Tracker against OpenAI's published documentation.

    O

    OnAirToday

    Newsroom

    OnAirTodayView the primary source

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

    GPT-5.2 is an OpenAI model for reasoning, coding, visual analysis and tool-assisted professional work. The benchmark figures below describe GPT-5.2 Thinking under OpenAI's published evaluation settings, not every API configuration; prices are USD per million tokens.

    What changed in this review:

    • SWE-bench Verified now reports OpenAI's published 80.0% Thinking result, rather than treating different evaluations as interchangeable.
    • Added OpenAI's published AIME 2025 and ARC-AGI-2 results.
    • Benchmark coverage now uses vendor-published Thinking results; earlier third-party and Elo figures are excluded.
    • GPT-5.3-Codex shipped as a later coding-focused family release; this profile remains specific to GPT-5.2.
    • Capability score 69 → 82

    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/api/pricing/
    • https://openai.com/index/introducing-gpt-5-3-codex/
    View the primary source →
    Share
    Filed under:AI Research

    Comments

    Keep comments relevant and respectful. No links, advertising, self-promotion, or spam.

    0/1000