Web frameworks compared

Four frameworks on three measures that share no unit: requests per second, memory and cold start on a logarithmic axis. One panel per measure, each with its own value axis, unit and verdict, over the same categories. On a phone the panels stack in one column. Illustrative values.

Web frameworks compared

examples/framework-benchmarks.json

{
  "schemaVersion": 1,
  "type": "multiples",
  "title": "Web frameworks compared",
  "source": "Illustrative values",
  "width": 960,
  "height": 420,
  "showValues": true,
  "columns": 3,
  "categories": [
    "Rust (Axum)",
    "Go (Gin)",
    "Node (Fastify)",
    "Python (FastAPI)"
  ],
  "panes": [
    {
      "title": "Requests per second",
      "note": "higher is better",
      "values": [
        118000,
        96000,
        64000,
        18000
      ],
      "valueAxis": {
        "unit": "req/s",
        "thousandsSeparator": true
      }
    },
    {
      "title": "Memory under load",
      "note": "lower is better",
      "values": [
        24,
        41,
        118,
        96
      ],
      "valueAxis": {
        "unit": "MB"
      }
    },
    {
      "title": "Cold start",
      "note": "lower is better",
      "values": [
        3,
        5,
        45,
        210
      ],
      "valueAxis": {
        "unit": "ms",
        "scale": "log"
      }
    }
  ],
  "mobile": {
    "width": 360,
    "height": 960,
    "columns": 1
  }
}

Source: Illustrative values

Show chart data

Data for Web frameworks compared
Category	Requests per second (req/s)	Memory under load (MB)	Cold start (ms)
Rust (Axum)	118,000	24	3
Go (Gin)	96,000	41	5
Node (Fastify)	64,000	118	45
Python (FastAPI)	18,000	96	210

Small multiples of bars (phone variant)
Compare candidates on unlike measures