benchmark-results / benchmark.py

Commit History

  1. Use configured bug and security floors and 70% semantic lint cutoff

    Verified signature
    b861a3e
  2. Add 50% floor results for current Perch question pack across three benchmarks (#20)⁠

    1add04d
  3. Update benchmark runners for current Perch scan questions (#17)⁠

    157d7a8

  1. Add benchmark adapters for released models (#11)⁠

    90964f9
  2. Publish runnable public benchmark evaluator (#10)⁠

    f459705