MCP-Universe (LLM w/ Function Calls): leaderboard
MCP-Universe benchmarks LLMs on application-centric multi-step workflows across 6 task categories. Function calling track from Salesforce AI Research.
Metric: Avg Evaluator Score. Source: mcp-universe.github.io. Saturation forecast: Around January 2027. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 61.67 | #66 |
| 2 | Grok 4.1 Fast | 60.58 | #217 |
| 3 | GPT-5 (Medium) | 59.29 | #98 (GPT-5) |
| 4 | Claude Sonnet 4.5 | 54.38 | #142 |
| 5 | Claude Sonnet 4 | 52.61 | #199 |
| 6 | Grok 4 Fast | 52.2 | #254 |
| 7 | Claude Sonnet 4 (Thinking) | 50.83 | #199 (Claude Sonnet 4) |
| 8 | Kimi K2 (Thinking) | 45.23 | #243 (Kimi K2) |
| 9 | Claude Haiku 4.5 | 44.63 | #280 |
| 10 | Grok Code Fast 1 | 43.89 | #337 |
| 11 | GPT-OSS-120B | 39.78 | #341 |
| 12 | GPT-4.1 | 39.58 | #248 |
| 13 | DeepSeek V3.1 | 38.88 | #264 |
| 14 | Kimi K2 0905 | 38.52 | #288 |
| 15 | GLM-4.5 | 37.58 | #275 |
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-function-calls · How It Works · Data refreshed daily, snapshot 2026-10-20.