List your company

ResearchFormerly WizWand

Software Engineering

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

16.1Accuracy

8

Feb 26, 2026

SWE-Bench Verified

76.2Success Rate

7

Jun 16, 2026

Terminal-Bench 2.0

75.1Public Score

7

Jun 12, 2026

SWE-Bench Verified

Claude-Opus-4.5

78.2SWE-Agent Score

7

Mar 4, 2026

SWE-Bench (val)

Claude 3.7 Sonnet

28.8Acc

7

Feb 26, 2026

SWE-bench Lite first 10 instances (test)

groq:llama-3.3-70b

0FCR

6

Jul 7, 2026

SWE-bench Verified (decidable subset)

Random rollout selection

286n_mix

6

Jun 23, 2026

SWE-Bench Lite 300-issue subset

32Accuracy

6

May 4, 2026

SWE-Bench Verified

ProRL Agent-14B (RL)

23.6Reproduced Success Rate

6

Mar 20, 2026

SWE-bench Verified

71Worst-Case Success Rate

6

Feb 26, 2026

73.7Score

5

Jun 23, 2026

81.9MTP Acceptance Rate

5

Jun 11, 2026

SWE-Bench Pro (held-out)

78Pass Rate

5

Jun 5, 2026

83.3WR (%)

5

May 21, 2026

72.4WR (%)

5

May 21, 2026

SWE-bench Lite (test)

62.7Resolve Rate

4

Jul 7, 2026

SWE tasks Configuration Integration n = 20 (test)

ParametricSkills

59.27F1 Score

4

Jun 30, 2026

SWE tasks Diagnosis Search n = 30 (test)

ParametricSkills

53.68F1 Score

4

Jun 30, 2026

SWE tasks Performance Data n = 10 (test)

49.64F1 Score

4

Jun 30, 2026

SWE tasks Refactor API Migration n = 39 (test)

ParametricSkills

56.12F1 Score

4

Jun 30, 2026

SWE tasks Robustness Fix n = 55 (test)

ParametricSkills

57.9F1 Score

4

Jun 30, 2026

SWE tasks Feature Patch n = 125 (test)

ParametricSkills

55.72F1 Score

4

Jun 30, 2026

21.4Productivity

4

Jun 15, 2026

SWE-bench 14 tasks

84.6Success Rate

4

May 19, 2026

SWE-Skills-Bench

Claude Sonnet 4.5

98Baseline Accuracy

4

May 13, 2026