.eval_results/swe_bench_verified.yaml

- dataset:
    id: SWE-bench/SWE-bench_Verified
    task_id: swe_bench_%_resolved
  value: 72.80
  source:
    url: https://www.swebench.com/
    name: SWE-Bench official evaluation
    user: nielsr
  notes: high reasoning, official

- dataset:
    id: SWE-bench/SWE-bench_Verified
    task_id: swe_bench_%_resolved
  value: 77.8
  source:
    url: https://huggingface.co/zai-org/GLM-5/
    name: Model card
    user: nielsr
  notes: Z.ai reported number