Benchmark
mcp-universe-llm
Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.
All agentic benchmarksScores from MCP-Universe.
- Models measured
- 23
- Scores recorded
- 24
- Sources
- 1
- IQ weight
- 2.00
Score distribution
- 24.1%
- Mean
- 22.9%
- Median
- 6.3%
- Std dev
- 18.2%
- 10th percentile
- 30.3%
- 90th percentile
- 14.3%
- Min
This benchmark has a weight of 2.00 for chat models, measured over 23 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | GPT 5high· MCP-Universe | 44.2% |
| 2 | Grok 4· MCP-Universe | 33.3% |
| 3 | Claude Sonnet 4thinking· MCP-Universe | 30.3% |
| 4 | Claude Opus 4.1· MCP-Universe | 29.4% |
| 5 | Claude Sonnet 4· MCP-Universe | 29.4% |
| 6 | Claude Opus 4· MCP-Universe | 28.1% |
| 7 | Grok 4 Fast· MCP-Universe | 27.3% |
| 8 | Grok Code Fast 1· MCP-Universe | 26.4% |
| 9 | GLM 4.6· MCP-Universe | 26% |
| 10 | GLM 4.5· MCP-Universe | 24.7% |
| 11 | Claude Sonnet 3.7· MCP-Universe | 24.2% |
| 12 | Qwen3 Coder· MCP-Universe | 22.9% |
| 13 | DeepSeek V3.1· MCP-Universe | 22.1% |
| 14 | Gemini 2.5 Pro· MCP-Universe | 22.1% |
| 15 | DeepSeek V3.1 Terminus· MCP-Universe | 21.7% |
| 16 | Gemini 2.5 Flash· MCP-Universe | 21.7% |
| 17 | DeepSeek V3.2 Exp· MCP-Universe | 19.9% |
| 18 | Kimi K2 0905· MCP-Universe | 19.9% |
| 19 | GLM 4.5 Air· MCP-Universe | 19.5% |
| 20 | Kimi K2· MCP-Universe | 19.1% |
| 21 | GPT 4.1· MCP-Universe | 18.2% |
| 22 | Qwen3 235B A22B Instruct 2507· MCP-Universe | 18.2% |
| 23 | GPT 4o· MCP-Universe | 15.6% |
| 24 | DeepSeek V3· MCP-Universe | 14.3% |