# NL2Repo-Bench — AI model rankings

Tests whether the AI can turn a natural-language requirement into working code across an entire repository, not just produce a single function or patch. Higher is better.

6 tracked models have a published NL2Repo-Bench score. Higher is better. Scores come from published lab reports and benchmark sources; recorded sources appear beside the scores.

## Ranking

| Rank | Model | Developer | Score | Source | Released |
| --- | --- | --- | --- | --- | --- |
| 1 | DeepSeek-V4.1-Flash | DeepSeek | 65.4% | Lab | Sep 10 2026 |
| 2 | Qwen3.8-Max-0902 | Qwen | 64.9% | Lab | Sep 2 2026 |
| 3 | GLM-5.3-Flash | Z.ai | 56.3% | Lab | Aug 26 2026 |
| 4 | Qwen3.8-Max | Qwen | 55.9% | Lab | Aug 3 2026 |
| 5 | Qwen3.8-Flash-Next | Qwen | 48.1% | Lab | Aug 26 2026 |
| 6 | Qwen3.8-27B | Qwen | 42.3% | Lab | Aug 14 2026 |


---

Canonical page: https://aireleasetracker.com/benchmark/nl2repo-bench
Site index: https://aireleasetracker.com/llms.txt
Source: AI Release Tracker (https://aireleasetracker.com). Most benchmark scores come from lab launch material; gathered results identify the leaderboard that published them.