MCPEvol-Bench - Task Fulfillment (Late Stage): leaderboard
Metric: Task Fulfillment score (1 to 10 rubric; servers after five 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 December 2026. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 | 6.77 |
| 2 | GPT-5.4 | 6.24 |
| 3 | Claude Sonnet 4.6 | 6.18 |
| 4 | Claude Sonnet 4.5 (Thinking) | 5.72 |
| 5 | GPT-5.1 | 5.32 |
| 6 | Gemma 4 31B (IT) | 5.16 |
| 7 | O3 | 5.07 |
| 8 | Gemini 2.5 Pro | 4.96 |
| 9 | GPT-4o | 4.43 |
| 10 | Qwen 3.5 27B | 4.13 |
| 11 | Llama 3.3 70B | 3.93 |
| 12 | Qwen 3.5 9B | 3.41 |
Interactive version: theaggregate.ai/benchmark?slug=mcpevol-bench-task-fulfillment-late-stage · How It Works · Data refreshed daily, snapshot 2026-09-29.