MCPBench
Createdmodelscope
Evaluates the performance of MCP servers for web search and database query tasks.
About
MCPBench is an evaluation framework designed to assess the performance of MCP servers, focusing on web search and database query functionalities. Compatible with both local and remote servers, it measures task completion accuracy, latency, and token consumption under consistent LLM and Agent configurations. The framework provides datasets and evaluation scripts to benchmark different MCP servers and is based on LangProBe: a Language Programs Benchmark.
Key Features
- Measures task completion accuracy
- 54 GitHub stars
- Supports local and remote MCP servers
- Evaluates database query MCP servers
- Evaluates web search MCP servers
- Provides datasets for evaluation
Use Cases
- Benchmarking database query MCP servers
- Benchmarking web search MCP servers
- Comparing performance of different MCP servers
