AgentCIBench: leaderboard
Metric: Leakage (self-reported). Source: benchmarklist.com. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 13.7 |
| 2 | GPT-5.4 | 18.8 |
| 3 | Claude Sonnet 4.6 | 46.2 |
| 4 | MiniMax-M2.7 | 58.1 |
| 5 | GPT-5.4 Mini | 60.7 |
| 6 | Kimi K2.6 | 62.4 |
| 7 | GPT-OSS-120B | 65.8 |
| 8 | Gemma 4 26B | 67.5 |
| 9 | Qwen 3.6 35B A3B | 76.9 |
| 10 | DeepSeek V4 Pro | 82.9 |
| 11 | GLM-5.1 | 85.5 |
| 12 | Grok 4.3 | 91.5 |
| 13 | Gemini 3 Flash | 93.2 |
| 14 | Qwen 3.6 Max | 97.4 |
| 15 | Gemini 3.1 Pro (Preview) | 98.3 |
Interactive version: theaggregate.ai/benchmark?slug=agentcibench · How It Works · Data refreshed daily, snapshot 2026-09-05.