AI Security

    Claude Sonnet 4.5: specified the published AIME result as AIME 2025

    Tracker verifies the benchmark edition cited in Anthropic’s documentation

    Specified the published AIME result as AIME 2025 — verified by the OnAirToday AI Progress Tracker against Anthropic's published documentation.

    O

    OnAirToday

    Newsroom

    OnAirToday

    The OnAirToday AI Progress Tracker recorded a verified change to Anthropic's Claude Sonnet 4.5 (version claude-sonnet-4-5-20250929): Specified the published AIME result as AIME 2025.

    Claude Sonnet 4.5 is an Anthropic model designed for coding, reasoning, computer use and multi-step agent workflows. Its published launch evaluations show improvements over Claude Sonnet 4, but factual errors, unreliable tool actions and susceptibility to malicious instructions remain important limitations.

    What changed in this review:

    • Added Anthropic's documented Claude Code cyberespionage misuse case without attributing it to an unconfirmed Sonnet version.
    • Capability score 67 → 74
    • Flaws and limitations updated
    • Risks and threats updated

    Anyone building on Claude Sonnet 4.5 should re-check the model's risk dashboard before shipping, since flaws and reported incidents are what moved in this review.

    The full profile — capabilities, benchmark citations, known flaws and risk guidance — is on the tracker at /ai-models/claude-sonnet-4-5.

    References:

    • https://www.anthropic.com/news/claude-sonnet-4-5
    • https://docs.anthropic.com/en/docs/about-claude/models/overview
    • https://docs.anthropic.com/en/docs/about-claude/pricing
    • https://www.anthropic.com/news/disrupting-AI-espionage
    AI Analysis

    Generated insight by OnAirToday AI

    Naming the exact AIME edition makes benchmark claims easier to interpret and compare. Verification against Anthropic’s documentation also gives readers a way to trace the reported result to its source.

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

    Comments

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

    0/1000