RuVerBench - Agentic Coding: leaderboard
Metric: Balanced Accuracy (%). Source: arxiv.org. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 (xHigh) | 89.4 |
| 2 | Gemini 3.1 Pro (Preview) (High) | 86.5 |
| 3 | Claude Opus 4.7 (Max) | 85 |
| 4 | Kimi K2.6 (Thinking) | 84.3 |
| 5 | DeepSeek V4 Pro (Max) | 83.8 |
| 6 | Claude Sonnet 4.6 (Max) | 83 |
| 7 | DeepSeek V4 Flash (Max) | 81.3 |
| 8 | GLM-5.1 (Thinking) | 80.8 |
| 9 | Qwen 3.6 Max (Preview) (Thinking) | 79.9 |
| 10 | MiMo V2.5 Pro (Thinking) | 78.5 |
| 11 | GPT-OSS-120B (High) | 77.1 |
| 12 | Qwen 3.5 122B A10B (Thinking) | 75 |
| 13 | Qwen 3.5 27B (Thinking) | 74 |
| 14 | GPT-OSS-20B (High) | 72.3 |
| 15 | Qwen 3.5 9B (Thinking) | 71.6 |
Interactive version: theaggregate.ai/benchmark?slug=ruverbench-agentic-coding · How It Works · Data refreshed daily, snapshot 2026-09-19.