GitHub

mcpmark

MCP 工具可观测性与评测

观察池 · 暂无正式排名

该项目当前未满足“两个有效维度 + 两种数据源”的主榜门槛。

— 未排名

可观测采用度缺失

—

动量当前有效 · 2026-09-12

关注度当前有效 · 2026-09-12

信号可信度 依据当前有数据的独立评分维度数量计算。

中2/3 · 1 种数据源

方法论 v2.0 · 快照 2026-09-12 · 超过 2 天视为过期

项目介绍

An evaluation suite for agentic models in real MCP tool environments (Notion / GitHub / Filesystem / Postgres / Playwright).

MCPMark provides a reproducible, extensible benchmark for researchers and engineers: one-command tasks, isolated sandboxes, auto-resume for failures, unified metrics, and aggregated reports.

🚀 MCPMark Verified is now the default. The standard tasks in this repository are the Verified set — every environment version-pinned and every verifier stabilized. Results from earlier task versions are deprecated and not directly comparable, so please report new numbers as MCPMark Verified. On the Verified set, gpt-5.5 (xhigh) leads at 92.9% and kimi-k2.7 reaches 81.1%. See…

各数据源

GitHub eval-sys/mcpmark 指标与趋势 →

460 Star

Star 460
Fork 48
提交 448
发布 4