NL2Repo — leaderboard
NL2Repo evaluates long-horizon coding capabilities including repository-level understanding, where models must generate or modify code across entire repositories from natural language specifications.
Metric: Score (self-reported). Source: benchmarklist.com. Status: saturation imminent. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.8 | 69.7 |
| 2 | GPT-5.5 | 50.7 |
| 3 | GLM-5.2 | 48.9 |
| 4 | Claude Opus 4.6 (Max) | 47.6 |
| 5 | Qwen 3.7 Max (Max) | 47.2 |
| 6 | GLM-5.1 | 42.7 |
| 7 | MiniMax-M3 | 42.1 |
| 8 | Qwen 3.7 Plus | 41.1 |
| 9 | DeepSeek V4 Pro | 35.5 |
| 10 | Qwen 3.6 Plus | 34.4 |
| 11 | Gemini 3.1 Pro (Preview) | 33.4 |
Interactive version: theaggregate.ai/benchmark?slug=nl2repo · How the rankings work · Data refreshed daily, snapshot 2026-07-22.