MCP-Universe (LLM w/ ReAct): leaderboard
MCP-Universe benchmarks LLMs on application-centric multi-step workflows: location navigation, repository management, financial analysis, 3D design, browser automation, and web search. ReAct prompting track.
Metric: Avg Evaluator Score. Source: mcp-universe.github.io. Saturation forecast: Around December 2027. 28 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 (High) | 62.82 | #98 (GPT-5) |
| 2 | GPT-5 (Medium) | 60.23 | #98 (GPT-5) |
| 3 | Claude Sonnet 4 (Thinking) | 51.74 | #199 (Claude Sonnet 4) |
| 4 | Claude Sonnet 4 | 50.61 | #199 |
| 5 | Claude Opus 4.1 | 49.14 | #140 |
| 6 | Grok 4 | 49.01 | #181 |
| 7 | Grok 4 Fast | 48.95 | #254 |
| 8 | Grok Code Fast 1 | 44.72 | #337 |
| 9 | GLM-4.6 | 43.94 | #252 |
| 10 | DeepSeek V3.1 | 43.23 | #264 |
| 11 | DeepSeek V3.1 Terminus | 42.68 | #220 |
| 12 | DeepSeek V3.2 Exp | 41.92 | #234 |
| 13 | Qwen 3 Coder 480B A35B Instruct | 41.39 | #315 |
| 14 | GPT-4.1 | 41.32 | #248 |
| 15 | Kimi K2 0905 | 41.28 | #288 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mcp-universe-llm-w-react · How It Works · Data refreshed daily, snapshot 2026-10-20.