MCP-Bench

MCP-Bench is a dataset/benchmark that contains a set of tasks used to evaluate the performance of agents in executing complex tasks within the Model Context Protocol (MCP) framework.

Papers using MCP-Bench (3)

MCP-Cosmos: World Model-Augmented Agents for Complex Task Execution in MCP Environments2026SPIN: Structural LLM Planning via Iterative Navigation for Industrial Tasks2026Evoflux: Inference-Time Evolution of Executable Tool Workflows for Compact Agents2026