AI Research

    GPT-5.4 Pro: model identifier, release date, context window, and pricing are no longer presented as confirmed specifications

    Tracker says key GPT-5.4 Pro details are not confirmed in OpenAI’s published documentation

    Model identifier, release date, context window, and pricing are no longer presented as confirmed specifications — verified by the OnAirToday AI Progress Tracker against OpenAI's published documentation.

    O

    OnAirToday

    Newsroom

    OnAirToday

    The OnAirToday AI Progress Tracker recorded a verified change to OpenAI's GPT-5.4 Pro (version gpt-5.4-pro): Model identifier, release date, context window, and pricing are no longer presented as confirmed specifications.

    GPT-5.4 Pro is identified as an OpenAI model, but confirmed specifications and benchmark results are not established. Separately reported sandbox escapes and a Hugging Face breach do not establish involvement by GPT-5.4 Pro.

    What changed in this review:

    • Unconfirmed GPQA Diamond, FrontierMath, and SimpleQA Verified scores are omitted.
    • Sandbox-escape and Hugging Face breach reports are recorded separately, without attributing either incident to GPT-5.4 Pro.

    Teams building on GPT-5.4 Pro 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-5-4-pro.

    References:

    • https://epoch.ai/benchmarks
    • https://thehackernews.com/2026/07/openai-says-its-own-ai-models-escaped.html
    AI Analysis

    Generated insight by OnAirToday AI

    This signals that key GPT-5.4 Pro details should not be treated as confirmed specifications. Readers and developers may need to wait for documentation that verifies them.

    Powered by AI • Not a substitute for human analysisMore Insights →
    Read the full story at OnAirToday →
    Share
    Filed under:AI Research

    Comments

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

    0/1000