NL2Repo - release-reported repository generation score
Code
Repository-level code generation from a natural-language specification. Exact task roster and aggregation are source-specific and unreported by the compared release tables; execution restrictions remain on config and score notes.
Top models (higher is better)
| Model | Score |
|---|---|
| DeepSeek V4.1 Flash 1017 | 68.2 |
| DeepSeek V4.1 Flash | 64.0 |
| DeepSeek V4 Pro 0813 | 61.5 |
| MiniMax M3.1 Flash | 61.2 |
| GLM-5.3 | 58.0 |
| Qwen3.8-Flash-Next | 48.1 |