RepoQA — leaderboard
RepoQA: Evaluates software-engineering agents on realistic issue resolution, repository navigation, testing, or maintenance workflows.
Metric: Score (self-reported). Source: benchmarklist.com. Status: saturated. 33 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 1.5 Pro | 90.6 |
| 2 | Claude 3 Opus | 90.6 |
| 3 | GPT-4o (2024-05-13) | 90.6 |
| 4 | Gemini 1.5 Flash | 90 |
| 5 | Claude 3 Sonnet (20240229) | 87.4 |
| 6 | DeepSeek V2 Chat | 83.4 |
| 7 | Llama 3 70B Instruct | 82.2 |
| 8 | Claude 3 Haiku | 81.8 |
| 9 | c4ai-command-r-plus | 78.4 |
| 10 | GPT-4 Turbo | 76.4 |
| 11 | Mixtral 8x22B Instruct (v0.1) | 67.8 |
| 12 | Qwen 1.5 72B Chat | 67 |
| 13 | Mistral 7B Instruct (v0.3) | 62 |
| 14 | GPT-3.5 Turbo | 60.4 |
| 15 | Llama 3 8B Instruct | 53.6 |
Interactive version: theaggregate.ai/benchmark?slug=repoqa · How the rankings work · Data refreshed daily, snapshot 2026-07-22.