Home / Frameworks / python / guillotina
HTTP server & framework benchmarks
guillotina benchmark results
guillotina 7.1 is a python web framework. At concurrency 64 it serves 10 987 requests per second with a P50 latency of 5.74 ms, ranking 323 of 383 overall and 30 of 46 among python frameworks.
Throughput, latency & reliability
All measured metrics for guillotina 7.1, by concurrency level.
Metric Concurrency 64 Concurrency 256 Concurrency 512
Requests / second 10 987 10 413 10 123
P50 latency 5.74 ms 24.12 ms 49.89 ms
P75 latency 5.96 ms 24.55 ms 50.36 ms
P90 latency 6.11 ms 24.96 ms 51.46 ms
P99 latency 6.44 ms 32.93 ms 82.14 ms
P99.999 latency 15.51 ms 569.90 ms 2884.61 ms
Average latency 5.82 ms 24.53 ms 50.40 ms
Minimum latency 1.03 ms 1.15 ms 1.16 ms
Standard deviation 0.45 ms 10.90 ms 76.44 ms
Total requests 164 815 156 209 151 881
Bytes received 38.98 MB 36.94 MB 35.92 MB
HTTP errors 0 0 0
Request timeouts 0 0 0
Socket connection errors 0 0 0
Socket read errors 0 0 0
Socket write errors 0 0 0
Position in this benchmark
Position of guillotina on requests per second.
Concurrency Overall Among python frameworks
64 323 of 383 30 of 46
256 325 of 383 30 of 46
512 325 of 383 31 of 46
Badge for your README
Show the rank of guillotina in your README. The image is updated together with the results.
[](https://web-frameworks-benchmark.vercel.app/frameworks/python/guillotina/)
[](https://web-frameworks-benchmark.vercel.app/frameworks/python/guillotina/)
Explore python frameworks
All 46 python frameworks