swebench-lite: by examples

Results Paper Code


Not solved by any model

There are 35 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-7746, django__django-11019, django__django-11564, django__django-11630, django__django-14667, django__django-14730, django__django-15695, django__django-16816, django__django-16820, matplotlib__matplotlib-22835, pallets__flask-5063, pydata__xarray-4493, pylint-dev__pylint-7228, scikit-learn__scikit-learn-11040, scikit-learn__scikit-learn-25638, sphinx-doc__sphinx-7686, sphinx-doc__sphinx-7738, sphinx-doc__sphinx-8282, sympy__sympy-11400, sympy__sympy-11870, sympy__sympy-12171, sympy__sympy-13146, sympy__sympy-14024, sympy__sympy-14308, sympy__sympy-14317, sympy__sympy-15308, sympy__sympy-16106, sympy__sympy-16281, sympy__sympy-17630, sympy__sympy-19254, sympy__sympy-20322, sympy__sympy-20639, sympy__sympy-21171, sympy__sympy-23191, sympy__sympy-24102

Problems solved by 1 model only

example_link model min_pass1_of_model
pytest-dev__pytest-5221 20250625_ExpeRepair-v1_claude-4-sonnet-20250514 0.603
django__django-14997 20250906_KGCompass_claude-4-sonnet-20250514 0.583
sympy__sympy-13895 20250906_KGCompass_claude-4-sonnet-20250514 0.583
sympy__sympy-18199 20250906_KGCompass_claude-4-sonnet-20250514 0.583
django__django-11905 20250906_KGCompass_claude-4-sonnet-20250514 0.583
sympy__sympy-12236 20250526_sweagent_claude-4-sonnet-20250514 0.567
django__django-16229 20250526_sweagent_claude-4-sonnet-20250514 0.567
sympy__sympy-13773 20250911_isea_claude-3.5-sonnet-20241022 0.513
sympy__sympy-11897 20250619_KGCompass_claude-3.5-sonnet-20241022 0.460
matplotlib__matplotlib-25433 20250901_entroPO_R2E_QwenCoder30BA3B 0.450
pydata__xarray-4248 20241207_kodu_sonnet_v1 0.447
django__django-15252 20241207_kodu_sonnet_v1 0.447
django__django-15738 20241207_kodu_sonnet_v1 0.447
sphinx-doc__sphinx-8273 20241207_kodu_sonnet_v1 0.447
sympy__sympy-13043 20241207_kodu_sonnet_v1 0.447
sympy__sympy-13437 20241207_kodu_sonnet_v1 0.447
astropy__astropy-14182 20240702_codestory_aide_mixed 0.430
django__django-13265 20241025_OpenHands-CodeAct-2.1-sonnet-20241022 0.417
django__django-13220 20250515_codartai 0.417
matplotlib__matplotlib-18869 20250515_codartai 0.417
matplotlib__matplotlib-25079 20250515_codartai 0.417
sphinx-doc__sphinx-8474 20250515_codartai 0.417
scikit-learn__scikit-learn-10949 20250515_codartai 0.417
scikit-learn__scikit-learn-10508 20250515_codartai 0.417
matplotlib__matplotlib-22711 20240627_abanteai_mentatbot_gpt4o 0.380
django__django-11742 20240627_abanteai_mentatbot_gpt4o 0.380
sympy__sympy-19007 20240622_Lingma_Agent 0.330
django__django-11910 20250207_aegis_o3mini 0.303
django__django-13768 20240523_aider 0.263

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
django__django-13768 0.012 -0.076
matplotlib__matplotlib-23299 0.036 -0.074
sympy__sympy-24909 0.107 -0.038
django__django-11910 0.012 -0.028
sympy__sympy-19007 0.012 -0.002
sympy__sympy-18835 0.036 0.013
matplotlib__matplotlib-22711 0.012 0.032
django__django-11742 0.012 0.032
pytest-dev__pytest-8365 0.071 0.038
pylint-dev__pylint-6506 0.155 0.047

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.