# MCPMark

> Stress-testing benchmark for agentic models using real MCP environments (Notion, GitHub, Filesystem, Postgres, Playwright), with strict automated verifiers and sandboxed tasks.

- id: `mcpmark`
- category: Information (`information`)
- url: https://mcpmark.ai
- repo: https://github.com/eval-sys/mcpmark
- docs: https://mcpmark.ai/docs
- status: active
- pricing: open-source
- license: Apache-2.0
- tags: benchmark, mcp, agent-evaluation, leaderboard, tool-use
- related: MCP-Universe (https://indexagentica.com/entries/mcp-universe/); Berkeley Function Calling Leaderboard (BFCL) (https://indexagentica.com/entries/bfcl/); Terminal-Bench (https://indexagentica.com/entries/terminal-bench/); WebArena (https://indexagentica.com/entries/webarena/)
- sources: https://github.com/eval-sys/mcpmark , https://mcpmark.ai , https://arxiv.org/abs/2509.24002 , https://huggingface.co/datasets/Jakumetsu/mcpmark-trajectory-log
- added: 2026-10-18
- updated: 2026-10-18
- last verified: 2026-10-18
- featured in: Agent benchmarks compared (https://indexagentica.com/compare/agent-benchmarks/)
- maintainer: eval-sys
- submitted by: agentica-curator
- page: https://indexagentica.com/entries/mcpmark/
- json: https://indexagentica.com/api/entries/mcpmark.json

One-command tasks, isolated sandboxes, auto-resume and aggregated reports. Since June 2026 the default task set is MCPMark Verified (version-pinned environments, stabilized verifiers); earlier results are not directly comparable. Trajectory logs are published on Hugging Face. Paper: arXiv 2509.24002.