swebench-verified: by examples

Results Paper Code


Not solved by any model

There are 34 examples not solved by any model. Solving some of these can be a good signal that your model is indeed better than leading models if these are good problems.
astropy__astropy-13398, astropy__astropy-13977, django__django-10554, django__django-10999, django__django-11087, django__django-11400, django__django-12406, django__django-13212, django__django-13513, django__django-14011, django__django-14034, django__django-14155, django__django-15629, django__django-16263, django__django-16502, django__django-16631, django__django-16667, matplotlib__matplotlib-25479, matplotlib__matplotlib-26466, pydata__xarray-6992, pydata__xarray-7229, pylint-dev__pylint-4551, pylint-dev__pylint-4604, pylint-dev__pylint-4661, pytest-dev__pytest-10356, sphinx-doc__sphinx-11510, sphinx-doc__sphinx-7462, sphinx-doc__sphinx-7748, sphinx-doc__sphinx-9461, sympy__sympy-13852, sympy__sympy-16597, sympy__sympy-20428, sympy__sympy-20438, sympy__sympy-21930

Problems solved by 1 model only

example_link model min_pass1_of_model
django__django-11820 20250804_epam-ai-run-claude-4-sonnet 0.768
django__django-15252 20250901_warp 0.756
sphinx-doc__sphinx-7590 20250915_JoyCode 0.746
sympy__sympy-18199 20250915_JoyCode 0.746
sympy__sympy-21596 20250915_JoyCode 0.746
sympy__sympy-17630 20251015_Prometheus_v1.2.1_gpt5 0.744
sphinx-doc__sphinx-9229 20251015_Prometheus_v1.2.1_gpt5 0.744
matplotlib__matplotlib-21568 20250623_warp 0.710
django__django-14792 20250805_openhands-Qwen3-Coder-480B-A35B-Instruct 0.696

Suspect problems

These are 10 problems with the lowest correlation with the overall evaluation (i.e. better models tend to do worse on these. )

example_link pass1_of_ex tau
astropy__astropy-7606 0.053 -0.200
sympy__sympy-18763 0.412 -0.198
django__django-11790 0.214 -0.175
sphinx-doc__sphinx-10614 0.015 -0.046
matplotlib__matplotlib-23476 0.015 -0.043
scikit-learn__scikit-learn-26194 0.023 -0.040
django__django-11477 0.015 -0.016
sympy__sympy-17318 0.076 -0.005
django__django-13794 0.053 0.001
astropy__astropy-13033 0.015 0.004

Histogram of accuracies

Histogram of problems by the accuracy on each problem.

Histogram of difficulties

Histogram of problems by the minimum win rate to solve each problem.