LLM Stats (MCP Atlas): leaderboard
Evaluates tool-use capability via Model Context Protocol tasks spanning real MCP servers and tools, testing multi-step workflow execution.
Metric: Score (%). Source: llm-stats.com. Source: qwen.ai. Saturation forecast: Around January 2027. 38 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Muse Spark Open | 89.2 | #54 |
| 2 | Muse Spark 1.1 | 88.1 | #52 |
| 3 | Kimi K3 | 84.2 | #30 |
| 4 | Seed 2.1 Pro | 83.8 | #67 |
| 5 | Hy4 preview | 83.7 | #38 |
| 6 | Gemini 3.5 Flash | 83.6 | #60 |
| 7 | Claude Opus 4.8 | 82.2 | #37 |
| 8 | Seed 2.1 Turbo | 80.3 | #86 |
| 9 | Inkling Small | 79.6 | #138 |
| 10 | Hy3 | 79.1 | #93 |
| 11 | Claude Opus 4.7 | 77.3 | #49 |
| 12 | GLM-5.2 | 76.8 | #73 |
| 13 | Qwen 3.7 Max | 76.4 | #75 |
| 14 | Kimi K2.7 Code | 76 | #89 |
| 15 | Inkling | 76 | #126 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-mcp-atlas · How It Works · Data refreshed daily, snapshot 2026-10-20.