MCPEvol-Bench - Planning Effectiveness (Middle Stage): leaderboard

Metric: Planning Effectiveness score (1 to 10 rubric; servers after three evolution rounds; 201 synthesized multi-server tasks over a fixed set of candidate MCP servers from 123 real servers; rubric scores on a 1 to 10 scale from a DeepSeek-Chat judge over the instruction, final answer and execution trajectory (MCP-Bench evaluation framework)). Source: arxiv.org. Saturation forecast: Around April 2027. 12 models tracked.

Top models

#ModelScore
1Gemma 4 31B (IT)6.69
2Claude Sonnet 4.66.42
3Claude Opus 4.66.18
4Qwen 3.5 27B5.97
5Claude Sonnet 4.5 (Thinking)5.56
6GPT-5.45.52
7Qwen 3.5 9B4.45
8Gemini 2.5 Pro4.03
9Llama 3.3 70B3.74
10GPT-5.13.43
11O33.21
12GPT-4o2.74

Interactive version: theaggregate.ai/benchmark?slug=mcpevol-bench-planning-effectiveness-middle-stage · How It Works · Data refreshed daily, snapshot 2026-09-29.