Everyone Is Benchmarking MCP Servers Wrong
Existing MCP benchmarks rank models, not servers. Here's how to A/B test whether your MCP server actually improves agent performance.
3 items
Existing MCP benchmarks rank models, not servers. Here's how to A/B test whether your MCP server actually improves agent performance.
MCP developers are shipping tools without evidence they work. I built mcpbr to find out. Here are results from a 500-task controlled SWE-bench experiment that surprised us.
Do MCP tools serialize in Claude Code? Empirical study: readOnlyHint controls parallelism, IPC overhead is ~5ms/call.