MCPBench Evaluation Framework

MCPOfficialOpen SourceApache-2.067.7

by modelscope • Uncategorized

MCPBench is a benchmark framework for evaluating the performance of MCP servers across Web Search, Database Query, and GAIA tasks.

Example Use Cases

1
Evaluate and compare MCP server performance on Web Search tasks.
2
Benchmark MCP servers on Database Query tasks.
3
Assess MCP server capabilities on GAIA tasks.

Description

MCPBench provides an evaluation framework compatible with both local and remote MCP servers, assessing their task completion accuracy, latency, and token consumption under consistent LLM and agent configurations. It supports multiple MCP server types including Web Search, Database Query, and GAIA, and offers datasets and scripts to facilitate benchmarking. The framework helps users compare different MCP servers like Brave Search and DuckDuckGo to identify the best performing solutions for their needs.

Quick Actions

View on GitHub

Security

Scanned 1 month(s) ago

Risk Level

MINIMAL

Read-only data retrieval, no side effects

Trust Score

D42/100

6/17 checks passed

Scores are informational only and provided “as is” without warranty. AgentHotspot assumes no liability for actions taken based on these ratings.

Quick Stats

Service TypeMCP

Pricing ModelFree

Capabilities0 Tools / 0 Prompts / 0 Resources

Ownermodelscope

CategoryUncategorized

DependenciesStandalone