Skip to content

Benchmark

MCP-Universe

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages. Higher is better.

Models measured
1
Scores recorded
3
Sources
2
IQ weight
No weight published

Score distribution

43.5%
Mean
45.9%
Median
3.4%
Std dev
38.6%
10th percentile
45.9%
90th percentile
38.6%
Min
38.6%45.9%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1DeepSeek V3.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim45.9%
2DeepSeek V3.2Thinking mode · temperature 1.0 · 128K context · internal search/Playwright environment · tool-role outputs· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.45.9%
3DeepSeek V3.2Non-thinking mode · temperature 1.0 · 128K context · internal search/Playwright environment · tool-role outputs· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.38.6%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Data by LLM Stats