NL2RepoBrowse 284 families
NL2Repo
Can the model complete substantial programming work under this benchmark's agent setup?
Version / measure
- Domain
- Coding
- Ranked models
- 13
- Reported values
- 16
- Reports
- 3 · lab
- Direction
- Higher is better
- Status
- Version not specified
Rankings
Best reported score per model across effort settings
| # | Model | Best setting | Score |
|---|---|---|---|
| 1 | Claude Opus 4.8Anthropic | reportedcost n/a1 value · Jul 31 · source | 69.7% |
| 2 | DeepSeek V4 Flash 0731DeepSeekOW | maxcost n/a1 value · Jul 31 · source | 54.2% |
| 3 | GLM-5.2Z.aiOW | reportedcost n/a1 value · Jul 31 · source | 48.9% |
| 4 | Claude Opus 4.5Anthropic | reportedcost n/a1 value · Apr 21 · source | 43.2% |
| 5 | DeepSeek-V4-FlashDeepSeekOW | reportedcost n/a1 value · Jul 31 · source | 39.4% |
| 6 | DeepSeek V4 ProDeepSeekOW | reportedcost n/a1 value · Jul 31 · source | 38.5% |
| 7 | Qwen3.6-27BQwen | reportedcost n/a1 value · Apr 21 · source | 36.2% |
| 8 | Qwen3.5-397B-A17BQwen | reportedcost n/a1 value · Apr 21 · source | 32.2% |
| 9 | Qwen3.6-35B-A3BQwen | reportedcost n/a2 values · Apr 21 · source | 29.4% |
| 10 | Qwen3.5-27BQwen | reportedcost n/a2 values · Apr 21 · source | 27.3% |
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.
Other models evaluated via Claude Code with temp=1.0, top_p=0.95, max_turns=900.
Method / source ↗