mcp-server-bench
Collection
This is a collection of Benchmarking results between Gradio and FastMCP • 4 items • Updated
scenario_id large_stringlengths 38 55 | server large_stringclasses 2
values | protocol large_stringclasses 1
value | tool large_stringclasses 2
values | virtual_users int64 1 10 | concurrency_limit float64 1 1 ⌀ | duration_s float64 0 30.4 | total_requests int64 0 4.55k | successful_requests int64 0 4.55k | failed_requests int64 0 0 | error_rate_pct float64 0 0 | throughput_rps float64 0 151 | latency_min_ms float64 0 4.13k | latency_p50_ms float64 0 4.15k | latency_p75_ms float64 0 4.16k | latency_p90_ms float64 0 8.35k | latency_p95_ms float64 0 8.98k | latency_p99_ms float64 0 9.02k | latency_max_ms float64 0 9.03k | latency_mean_ms float64 0 4.66k | latency_stddev_ms float64 0 1.54k | avg_cpu_pct float64 0 0 | avg_memory_mb float64 0 5.68 | peak_memory_mb float64 0 5.68 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
fastmcp__mcp_streamable__echo__vu1__clunlimited | fastmcp | mcp_streamable | echo | 1 | null | 29.98 | 3,785 | 3,785 | 0 | 0 | 126.23 | 0 | 7.506 | 15.625 | 15.633 | 15.653 | 19.636 | 35.151 | 7.849 | 7.636 | 0 | 5.64 | 5.64 |
fastmcp__mcp_streamable__echo__vu10__clunlimited | fastmcp | mcp_streamable | echo | 10 | null | 30.03 | 2,253 | 2,253 | 0 | 0 | 75.02 | 35.458 | 115.045 | 157.137 | 203.987 | 239.743 | 322.767 | 458.957 | 130.394 | 57.577 | 0 | 5.65 | 5.67 |
fastmcp__mcp_streamable__async_sleep__vu1__clunlimited | fastmcp | mcp_streamable | async_sleep | 1 | null | 29.99 | 526 | 526 | 0 | 0 | 17.54 | 44.346 | 62.46 | 63.013 | 63.625 | 66.519 | 78.367 | 79.935 | 57.101 | 8.36 | 0 | 5.67 | 5.67 |
fastmcp__mcp_streamable__async_sleep__vu10__clunlimited | fastmcp | mcp_streamable | async_sleep | 10 | null | 30.41 | 1,334 | 1,334 | 0 | 0 | 43.87 | 55.091 | 146.304 | 295.423 | 477.027 | 606.654 | 913.331 | 1,516.83 | 222.449 | 188.637 | 0 | 5.66 | 5.66 |
gradio__mcp_streamable__echo__vu1__cl1 | gradio | mcp_streamable | echo | 1 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
gradio__mcp_streamable__echo__vu1__cl1__noq | gradio | mcp_streamable | echo | 1 | 1 | 30.01 | 1,629 | 1,629 | 0 | 0 | 54.28 | 0 | 16.017 | 25.431 | 31.357 | 34.984 | 43.468 | 167.634 | 18.311 | 11.397 | 0 | 5.68 | 5.68 |
gradio__mcp_streamable__echo__vu10__cl1 | gradio | mcp_streamable | echo | 10 | 1 | 28.96 | 80 | 80 | 0 | 0 | 2.76 | 3,426.991 | 4,123.11 | 4,146.098 | 8,349.039 | 8,977.496 | 9,016.052 | 9,032.483 | 4,650.837 | 1,535.19 | 0 | 5.66 | 5.66 |
gradio__mcp_streamable__echo__vu10__cl1__noq | gradio | mcp_streamable | echo | 10 | 1 | 30.04 | 4,549 | 4,549 | 0 | 0 | 151.45 | 12.207 | 62.953 | 73.506 | 94.658 | 110.314 | 146.306 | 346.802 | 64.663 | 23.449 | 0 | 5.66 | 5.66 |
gradio__mcp_streamable__async_sleep__vu1__cl1 | gradio | mcp_streamable | async_sleep | 1 | 1 | 29.04 | 8 | 8 | 0 | 0 | 0.28 | 4,134.819 | 4,149.33 | 4,158.915 | 5,388.08 | 6,820.616 | 7,966.644 | 8,253.152 | 4,661.271 | 1,357.627 | 0 | 5.66 | 5.66 |
gradio__mcp_streamable__async_sleep__vu1__cl1__noq | gradio | mcp_streamable | async_sleep | 1 | 1 | 29.99 | 497 | 497 | 0 | 0 | 16.57 | 31.29 | 62.939 | 63.712 | 67.735 | 75.147 | 79.91 | 83.831 | 60.445 | 8.635 | 0 | 5.66 | 5.66 |
gradio__mcp_streamable__async_sleep__vu10__cl1 | gradio | mcp_streamable | async_sleep | 10 | 1 | 28.9 | 80 | 80 | 0 | 0 | 2.77 | 3,784.411 | 4,128.612 | 4,146.769 | 8,222.441 | 8,304.213 | 8,599.548 | 8,606.705 | 4,637.582 | 1,409.364 | 0 | 5.66 | 5.66 |
gradio__mcp_streamable__async_sleep__vu10__cl1__noq | gradio | mcp_streamable | async_sleep | 10 | 1 | 30.01 | 3,798 | 3,798 | 0 | 0 | 126.56 | 44.026 | 78.523 | 85.501 | 95.979 | 109.571 | 125.906 | 165.462 | 77.578 | 15.93 | 0 | 5.66 | 5.66 |
🔬 Gradio vs FastMCP Benchmark Report
Generated: 2026-03-04T10:26:47.242725
Total scenarios: 12
Executive Summary
- echo: Gradio wins (151.4 vs 126.2 RPS, 1.2x difference)
- Gradio best config: concurrency_limit=1.0
- async_sleep: Gradio wins (126.6 vs 43.9 RPS, 2.88x difference)
- Gradio best config: concurrency_limit=1.0
Throughput (Requests/Second)
| ('fastmcp', 'mcp_streamable') | ('gradio', 'mcp_streamable') | |
|---|---|---|
| ('async_sleep', 1) | 17.54 | 16.57 |
| ('async_sleep', 10) | 43.87 | 126.56 |
| ('echo', 1) | 126.23 | 54.28 |
| ('echo', 10) | 75.02 | 151.45 |
Latency p50 (ms)
| ('fastmcp', 'mcp_streamable') | ('gradio', 'mcp_streamable') | |
|---|---|---|
| ('async_sleep', 1) | 62.46 | 62.939 |
| ('async_sleep', 10) | 146.304 | 78.523 |
| ('echo', 1) | 7.506 | 0 |
| ('echo', 10) | 115.045 | 62.953 |
Gradio concurrency_limit Scaling
How does Gradio's throughput change as concurrency_limit increases?
| 1.0 | |
|---|---|
| ('async_sleep', 1) | 8.425 |
| ('async_sleep', 10) | 64.665 |
| ('echo', 1) | 27.14 |
| ('echo', 10) | 77.105 |
Protocol Overhead: HTTP API vs MCP
Comparing latency of the same tool called via REST API vs MCP protocol:
| mcp_streamable | |
|---|---|
| ('fastmcp', 'async_sleep') | 104.382 |
| ('fastmcp', 'echo') | 61.2755 |
| ('gradio', 'async_sleep') | 2104.85 |
| ('gradio', 'echo') | 1050.52 |
Error Rates
All scenarios completed with 0% error rate. ✅
Resource Usage
| server | ('mean', 'avg_cpu_pct') | ('mean', 'avg_memory_mb') | ('mean', 'peak_memory_mb') | ('max', 'avg_cpu_pct') | ('max', 'avg_memory_mb') | ('max', 'peak_memory_mb') |
|---|---|---|---|---|---|---|
| fastmcp | 0 | 5.655 | 5.66 | 0 | 5.67 | 5.67 |
| gradio | 0 | 4.955 | 4.955 | 0 | 5.68 | 5.68 |
Benchmark Charts
Methodology
- Both servers use identical tool implementations (imported from shared_tools.py)
- Each scenario runs in an isolated server subprocess
- Warmup period excluded from measurements
- Load generated by async httpx workers (not external tools)
- MCP tests use full protocol lifecycle (initialize → call_tool)
- System metrics sampled every 1s via psutil
Benchmarks generated by mcp-server-bench
- Downloads last month
- 86