GitHub

Model Evaluation and Threat Research (METR)

METR is a research nonprofit that works on assessing whether cutting-edge AI systems could pose catastrophic risks to society.

We build the science of accurately assessing risks, so that humanity is informed before developing transformative AI systems.

Read more about our work here.

Our Software

Pinned Loading

  1. Run Inspect AI evals in the cloud

    Python 51 34

Repositories

Showing 10 of 63 repositories

  • hawk Public

    Run Inspect AI evals in the cloud

    METR/hawk's past year of commit activity

  • METR/inspect_ai's past year of commit activity

    Python

    9

    MIT

    659 0 0

    Updated Aug 17, 2026

  • METR/inspect_scout's past year of commit activity

    Python

    1

    MIT

    33 0 2

    Updated Aug 17, 2026

  • METR/ts-mono's past year of commit activity

    TypeScript

    0 20 0 0

    Updated Aug 17, 2026

  • METR/inspect_k8s_sandbox's past year of commit activity

    Python

    2

    MIT

    28 0 0

    Updated Aug 15, 2026

  • inspect-agents Public

    A collection of METR wrappers around Inspect agents and of METR scanners for Inspect Scout. Intended to allow consistent usage and customization.

    METR/inspect-agents's past year of commit activity

    Python

    8

    MIT

    1 6 11

    Updated Aug 15, 2026

  • METR/inspect-test-utils's past year of commit activity

    Python

    0

    MIT

    2 1 1

    Updated Aug 13, 2026

  • bombe-engine Public

    METR fork of CharlieBrej/Bombe: the Bombe puzzle engine, assets removed, with eval-specific modifications.

    METR/bombe-engine's past year of commit activity

    C++

    0

    CC-BY-4.0

    0 0 0

    Updated Aug 2, 2026

  • METR/hcast-public's past year of commit activity

    HTML

    22 5 3 1

    Updated Jul 28, 2026

  • METR/public-tasks's past year of commit activity

    HTML

    129 19 1 2

    Updated Jul 28, 2026

Read the original on github.com ↗