# MCP-Bench

> Accenture's benchmark for tool-using agents on complex real-world tasks across 28 live MCP servers and 250 tools (finance, travel, science, academic search).

- id: `mcp-bench`
- category: Information (`information`)
- url: https://arxiv.org/abs/2508.20453
- repo: https://github.com/Accenture/mcp-bench
- status: active
- pricing: free
- license: Apache-2.0
- tags: benchmark, mcp, tool-use, agents, evals
- related: Toolathlon (Tool Decathlon) (https://indexagentica.com/entries/toolathlon/); MCPMark (https://indexagentica.com/entries/mcpmark/); MCP-Universe (https://indexagentica.com/entries/mcp-universe/); Berkeley Function Calling Leaderboard (BFCL) (https://indexagentica.com/entries/bfcl/)
- sources: https://arxiv.org/abs/2508.20453 , https://github.com/Accenture/mcp-bench , https://huggingface.co/spaces/mcpbench/mcp-bench
- added: 2026-10-20
- updated: 2026-10-20
- last verified: 2026-10-20
- maintainer: Accenture
- submitted by: agentica-curator
- page: https://indexagentica.com/entries/mcp-bench/
- json: https://indexagentica.com/api/entries/mcp-bench.json

Tests cross-tool coordination, precise parameter control and planning, end to end over MCP: the repo ships the 28 server implementations, task synthesis and the runner (`run_benchmark.py`), and a Hugging Face Space hosts the leaderboard. Accepted to the NeurIPS 2025 Workshop on Scaling Environments for Agents (arXiv 2508.20453). The README badge says Apache-2.0, but the repository has no license file.