MCPMark: A Benchmark for Stress-Testing Realistic and Comprehensive MCP Use
This addresses the need for more realistic and comprehensive benchmarks for MCP-based agents, which is crucial for developers and researchers evaluating LLM interactions with external systems, though it is incremental as it builds on existing MCP benchmarking efforts.
The paper tackles the problem of narrow and unrealistic benchmarks for MCP (Model Context Protocol) use by introducing MCPMark, a benchmark with 127 tasks that demand diverse CRUD operations, and finds that even top models like gpt-5-medium achieve only 52.56% pass@1 and 33.86% pass^4, with LLMs requiring 16.2 execution turns and 17.4 tool calls per task on average.
MCP standardizes how LLMs interact with external systems, forming the foundation for general agents. However, existing MCP benchmarks remain narrow in scope: they focus on read-heavy tasks or tasks with limited interaction depth, and fail to capture the complexity and realism of real-world workflows. To address this gap, we propose MCPMark, a benchmark designed to evaluate MCP use in a more realistic and comprehensive manner. It consists of $127$ high-quality tasks collaboratively created by domain experts and AI agents. Each task begins with a curated initial state and includes a programmatic script for automatic verification. These tasks demand richer and more diverse interactions with the environment, involving a broad range of create, read, update, and delete (CRUD) operations. We conduct a comprehensive evaluation of cutting-edge LLMs using a minimal agent framework that operates in a tool-calling loop. Empirical results show that the best-performing model, gpt-5-medium, reaches only $52.56$\% pass@1 and $33.86$\% pass^4, while other widely regarded strong models, including claude-sonnet-4 and o3, fall below $30$\% pass@1 and $15$\% pass^4. On average, LLMs require $16.2$ execution turns and $17.4$ tool calls per task, significantly surpassing those in previous MCP benchmarks and highlighting the stress-testing nature of MCPMark.