ComplexMCP: leaderboard
Current LLM agents are proficient at calling isolated APIs but struggle with the "last mile" of commercial software automation. In real-world scenarios, tools are not independent; they are atomic, interdependent, and prone to environmental noise. We introduce ComplexMCP, a benchmark designed to evaluate agents in these rigorous conditions. Built on the Model Context Protocol (MCP), $
Metric: Success Rate (self-reported). Source: benchmarklist.com. Saturation forecast: Around February 2027. 17 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash | 55.31 | #99 |
| 2 | Gemini 3 Pro | 44.67 | #80 |
| 3 | GLM-4.7 | 42.55 | #192 |
| 4 | Claude Opus 4 | 41.84 | #163 |
| 5 | Claude Sonnet 4.5 | 39.71 | #142 |
| 6 | Claude Sonnet 4 | 38.29 | #199 |
| 7 | Qwen 3 Max | 31.2 | #208 |
| 8 | Llama 3.1 405B Instruct | 26.94 | #457 |
| 9 | Claude 3.5 Sonnet | 26.26 | #350 |
| 10 | Kimi K2 | 26.22 | #243 |
| 11 | Gemini 2.5 Pro | 24.81 | #152 |
| 12 | Llama 3.3 70B Instruct | 21.89 | #534 |
| 13 | DeepSeek V3 | 19.86 | #321 |
| 14 | GPT-5.1 | 19.14 | #135 |
| 15 | GPT-4o | 14.89 | #343 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=complexmcp · How It Works · Data refreshed daily, snapshot 2026-10-20.