Skip to content

@METR METR

Sources

    Repositories list

    39 repositories

    • hawk

      Public
      Run Inspect AI evals in the cloud
      awsllmevals
      awsllmevalsinspect-ai
      Python
      •
      MIT License
      •51•87•146•170•Updated Oct 6, 2026Oct 6, 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.
      Python
      •
      MIT License
      •4•10•2•21•Updated Oct 5, 2026Oct 5, 2026

    • inspect-eagle

      Public
      macOS GUI to view large inspect samples. Integrated with Hawk
      Swift
      •0•0•0•0•Updated Sep 28, 2026Sep 28, 2026

    • inspect-test-utils

      Public
      Utilities for testing Inspect AI
      Python
      •
      MIT License
      •2•1•1•0•Updated Sep 22, 2026Sep 22, 2026

    • inspect-tasks-public

      Public
      Python
      •
      Other
      •9•4•0•6•Updated Sep 9, 2026Sep 9, 2026

    • inspect-metr-task-bridge

      Public
      Python
      •2•1•1•4•Updated Sep 3, 2026Sep 3, 2026

    • inspect-eval-utils

      Public
      Python
      •
      MIT License
      •0•0•0•0•Updated Aug 27, 2026Aug 27, 2026

    • task-assets

      Public
      Python
      •0•0•1•1•Updated Aug 19, 2026Aug 19, 2026

    • bombe-engine

      Public
      METR fork of CharlieBrej/Bombe: the Bombe puzzle engine, assets removed, with eval-specific modifications.
      C++
      •
      Creative Commons Attribution 4.0 International
      •0•0•0•0•Updated Aug 2, 2026Aug 2, 2026

    • hcast-public

      Public
      HTML
      •
      Other
      •5•24•3•1•Updated Jul 28, 2026Jul 28, 2026

    • public-tasks

      Public
      HTML
      •
      Other
      •22•134•7•5•Updated Jul 28, 2026Jul 28, 2026

    • triframe_inspect

      Public
      Python
      •0•4•0•0•Updated Jul 16, 2026Jul 16, 2026

    • .github

      Public
      0•0•0•0•Updated Jul 8, 2026Jul 8, 2026

    • local_manual_scoring_tool

      Public
      Python
      •0•0•0•1•Updated Jul 7, 2026Jul 7, 2026

    • vivaria

      Public
      Vivaria is METR's tool for running evaluations and conducting agent elicitation research.
      aielicitationai-evaluation
      aielicitationai-evaluationevals
      TypeScript
      •
      MIT License
      •40•143•220•7•Updated May 18, 2026May 18, 2026

    • evaluation-resources

      Public
      SCSS
      •
      MIT License
      •4•3•0•2•Updated May 1, 2026May 1, 2026

    • inspect-tinker-bridge

      Public
      Inspect tasks <> Tinker RL envs
      Python
      •
      MIT License
      •4•8•0•0•Updated Mar 10, 2026Mar 10, 2026

    • eval-analysis-public

      Public
      Public repository containing METR's DVC pipeline for eval data analysis
      Python
      •55•335•11•4•Updated Mar 6, 2026Mar 6, 2026

    • Measuring-Late-2025-AI-on-OSS-Devs

      Public
      Python
      •2•8•0•1•Updated Feb 24, 2026Feb 24, 2026

    • Measuring-Early-2025-AI-on-Exp-OSS-Devs

      Public
      Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity: https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study…
      Python
      •2•17•0•1•Updated Feb 23, 2026Feb 23, 2026

    • modelscan-inspect

      Public
      Modelscan but in Inspect
      Python
      •0•2•0•0•Updated Jan 20, 2026Jan 20, 2026

    • inspect-verifiers-bridge

      Public
      Bridge for inspect <> verifiers.
      Python
      •
      MIT License
      •0•0•0•0•Updated Jan 7, 2026Jan 7, 2026

    • terraform-docker-build

      Public
      Build docker containers using docker build cloud without a docker daemon
      HCL
      •
      MIT License
      •0•1•0•0•Updated Jan 2, 2026Jan 2, 2026

    • cross-domain-horizon

      Public
      Estimate the time horizon of AIs over time on various domains like knowledge and vision
      Python
      •2•5•1•0•Updated Dec 3, 2025Dec 3, 2025

    • task-aux-vm-helpers

      Public
      Python
      •0•0•0•0•Updated Nov 5, 2025Nov 5, 2025

    • task-artifacts

      Public
      Python
      •0•0•0•0•Updated Nov 5, 2025Nov 5, 2025

    • RE-Bench

      Public
      Python
      •
      Other
      •24•169•6•1•Updated Oct 16, 2025Oct 16, 2025

    • task-protected-scoring

      Public
      Python
      •1•1•3•0•Updated Oct 6, 2025Oct 6, 2025

    • CoT-faithfulness-and-monitorability

      Public
      Python
      •1•3•1•1•Updated Sep 12, 2025Sep 12, 2025

    • swe-bench-labeling

      Public
      Some data on SWE-Bench diffs, for labeling.
      0•0•0•0•Updated Sep 3, 2025Sep 3, 2025
    ProTip! When viewing an organization's repositories, you can use the props. filter to filter by custom property.