SWE ProBrowse 284 families
SWE Pro
Can the model complete substantial programming work under this benchmark's agent setup?
Version / measure
- Domain
- Coding
- Ranked models
- 7
- Reported values
- 11
- Reports
- 1 · lab
- Direction
- Higher is better
- Status
- Version not specified
Rankings
Best reported score per model across effort settings
| # | Model | Best setting | Score |
|---|---|---|---|
| 1 | Kimi K2.6Moonshot AIOW | reportedcost n/a1 value · Apr 26 · source | 58.6% |
| 2 | GLM-5.1Z.aiOW | reportedcost n/a1 value · Apr 26 · source | 58.4% |
| 3 | GPT-5.4OpenAI | xhighcost n/a1 value · Apr 26 · source | 57.7% |
| 4 | Claude Opus 4.6Anthropic | maxcost n/a1 value · Apr 26 · source | 57.3% |
| 5 | DeepSeek V4 ProDeepSeekOW | maxcost n/a3 values · Apr 26 · source | 55.4% |
| 6 | Gemini 3.1 ProGoogle | highcost n/a1 value · Apr 26 · source | 54.2% |
| 7 | DeepSeek-V4-FlashDeepSeekOW | maxcost n/a3 values · Apr 26 · source | 52.6% |
Effort curve
Score vs. reported cost. Lines connect complete sweeps only.
No cost-linked effort sweep for this version.
Definition and comparison boundarypublic methodology
Can the model complete substantial programming work under this benchmark's agent setup?
Software implementation, debugging, or repository work under the published evaluation protocol. Higher is better. The value is the percentage reported in this lab's table.
Harness, tools, prompts, attempt count, and benchmark version can materially change the result. Compare within one reporting context.
Frontier comparison and across-modes table.
Method / source ↗