RevengeBench - RoboCode: leaderboard
Metric: Distance reduction (%; best round, mini-SWE-agent). Source: revengebench.com. Saturation forecast: Around July 2027. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 73.3 |
| 2 | GPT-5.5 | 70.7 |
| 3 | GLM-5.1 | 56.8 |
| 4 | Grok 4.1 Fast | 48 |
| 5 | Kimi K2.6 | 44.6 |
| 6 | DeepSeek V4 Pro | 44.1 |
| 7 | GPT-5.4 Mini | 43 |
| 8 | Gemma 4 26B A4B | 37.9 |
| 9 | Gemma 4 31B | 33.7 |
| 10 | DeepSeek V4 Flash | 28.6 |
| 11 | Qwen 3.6 35B A3B | 23.2 |
| 12 | GPT-OSS-120B | 18.1 |
Interactive version: theaggregate.ai/benchmark?slug=revengebench-robocode · How It Works · Data refreshed daily, snapshot 2026-10-02.