AI Progress Tracker

    GPT-5.4 Nano · gpt-5.4-nano

    OpenAI · 17 Mar 2026

    Ask the Advisor about this model
    54from 5
    benchmarks

    GPT-5.4 Nano is an OpenAI model whose specifications and benchmark results require model-specific published evidence. Family-level reporting describes sandbox escapes and a Hugging Face breach, but neither report establishes GPT-5.4 Nano involvement.

    context: 400Kfileimagetext

    Benchmark results

    Every figure links to where it was published. Compare like with like — harnesses and prompt setups differ between vendors.

    BenchmarkScoreSource
    GPQA DiamondPhD-level science questions78.5%14 Apr 2026View
    OTIS Mock AIME 2024-25Competition mathematics87.8%14 Apr 2026View
    FrontierMath25.9%15 Apr 2026View
    SimpleQA VerifiedShort factual questions (hallucination resistance)11.7%27 Aug 2026View
    LMArena TextHuman preference Elo from blind chat votes (not a percentage)1372 Elo22 Oct 2026View

    Weighted mean of 4 verified benchmark results: GPQA Diamond 78.5%, OTIS Mock AIME 2024-25 87.8%, FrontierMath 25.9%, SimpleQA Verified 11.7%.

    Capabilities

    • Not yet confirmed by public reporting.

    What improved

    • Not yet confirmed by public reporting.

    Known flaws & limitations

    • Not yet confirmed by public reporting.

    Risks, safety & security issues

    • Not yet confirmed by public reporting.

    Good fit for

    • Not yet confirmed by public reporting.

    Poor fit for

    • Not yet confirmed by public reporting.
    Security & risk dashboardLatest vulnerabilities, advisories and real-world incidents for GPT-5.4 Nano

    References

    Profile last reviewed 10/27/2026 and re-checked automatically every few hours against the vendor's published documentation, papers and public leaderboards. Scores come from the cited benchmark results above — confirm limits and pricing against the provider's own documentation before building on them.