Benchmark
SWE-bench-Live Verified
Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.
- Models measured
- 2
- Scores recorded
- 2
- Sources
- 1
- IQ weight
- No weight published
Score distribution
- 35.2%
- Mean
- 40%
- Median
- 4.8%
- Std dev
- 30.4%
- 10th percentile
- 40%
- 90th percentile
- 30.4%
- Min
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.5scaffold: swe-agent, set: verified, instances: 500, submitted: 2025-12-31· SWE-bench-Live leaderboard data | 40% |
| 2 | GPT 5scaffold: swe-agent, set: verified, instances: 500, submitted: 2025-12-31, effort: medium· SWE-bench-Live leaderboard data | 30.4% |