Morphos & Kosmesis are here.Explore
PraxisJS

Benchmarks

How PraxisJS compares to Solid, Svelte, Vue, Lit, Preact and React in js-framework-benchmark, with the full results table and methodology.

Benchmarks

PraxisJS is measured with js-framework-benchmark, the standard suite for comparing frontend frameworks. Every entry renders the same table of rows and runs the same operations in Chrome. Results below are from the keyed table: rows keep their identity, so a moved row is moved, not rewritten.

Overall: geometric mean of slowdowns
Each framework's time divided by the fastest entry, per benchmark, then averaged. 1.00 = fastest on every benchmark; bar length is the distance from 1.00. Lower is better.
vanillajs1.02x
PraxisJS1.03x
Solid1.14x
Svelte1.18x
Vue1.27x
Lit1.32x
Preact + signals1.37x
React hooks1.78x
PraxisJS runtime 0.8.0 · 1.03x the fastest entry, geometric mean of 9 benchmarks
PraxisJS other frameworks vanillajs (hand-written DOM)

How to read this

  • Lower is better. Every number is the median of the measured runs.
  • Total time is mostly the browser. A benchmark's duration runs from the click to the end of the next paint. Most of it is style, layout and paint, which costs the same for every framework that produces the same DOM. Framework differences live in the remaining few milliseconds, so small gaps in total time are real gaps in framework work.
  • vanillajs is hand-written DOM code, written for this benchmark only. It is the reference for what the browser alone costs, not a framework you would build an app with. It can be beaten when a framework does less DOM work than the hand-written version.
  • Slowdown (1.05x) is a framework's time divided by the fastest entry in that benchmark. The overall score is the geometric mean of those slowdowns.

Results

Duration: median per benchmark, with the slowdown vs the fastest entry
Benchmark
vanillajs
hand-written
PraxisJS
runtime 0.8.0
Solid
1.9.3
Svelte
5.42.1
Vue
3.5.39
Lit
3.2.0
Preact + signals
11
React hooks
19.2
create 1k
ms
24.2
1.00x
24.3
1.00x
26.9
1.11x
26.1
1.08x
28.2
1.17x
28.5
1.18x
32.3
1.33x
28.4
1.17x
replace 1k
ms
25.6
1.00x
27.2
1.06x
28.4
1.11x
28.9
1.13x
32.2
1.26x
30.8
1.20x
41.0
1.60x
34.5
1.35x
update every 10th
ms
14.0
1.03x
13.6
1.00x
17.5
1.29x
15.7
1.15x
17.6
1.29x
16.0
1.18x
14.9
1.10x
20.3
1.49x
select row
ms
4.5
1.15x
3.9
1.00x
4.6
1.18x
7.3
1.87x
5.6
1.44x
8.2
2.10x
5.8
1.49x
8.1
2.08x
swap rows
ms
15.9
1.01x
15.7
1.00x
17.3
1.10x
17.9
1.14x
18.3
1.17x
22.0
1.40x
20.7
1.32x
117.1
7.46x
remove row
ms
13.4
1.00x
13.7
1.02x
14.7
1.10x
14.2
1.06x
16.9
1.26x
15.4
1.15x
14.7
1.10x
15.8
1.18x
create 10k
ms
250.5
1.00x
258.7
1.03x
271.8
1.09x
273.9
1.09x
300.2
1.20x
293.2
1.17x
350.7
1.40x
444.6
1.77x
append 1k
ms
28.1
1.00x
29.0
1.03x
30.5
1.09x
30.1
1.07x
32.8
1.17x
34.0
1.21x
38.4
1.37x
34.2
1.22x
clear
ms
11.6
1.00x
12.6
1.09x
14.7
1.27x
14.1
1.22x
17.8
1.53x
18.1
1.56x
19.9
1.72x
22.6
1.95x
geometric mean1.021.031.141.181.271.321.371.78
Memory: median per benchmark, with the slowdown vs the fastest entry
Benchmark
vanillajs
hand-written
PraxisJS
runtime 0.8.0
Solid
1.9.3
Svelte
5.42.1
Vue
3.5.39
Lit
3.2.0
Preact + signals
11
React hooks
19.2
ready memory
MB
0.6
1.00x
0.7
1.23x
0.6
1.07x
0.7
1.18x
0.9
1.55x
0.7
1.27x
0.7
1.23x
1.2
2.11x
run memory
MB
1.9
1.00x
2.7
1.46x
2.7
1.43x
2.9
1.54x
3.9
2.08x
2.8
1.50x
4.9
2.62x
4.4
2.35x
create/clear 1k x5
MB
0.6
1.00x
0.9
1.44x
0.8
1.20x
1.0
1.53x
1.2
1.86x
0.9
1.38x
9.0
14.03x
2.0
3.09x
Startup and size: median per benchmark, with the slowdown vs the fastest entry
Benchmark
vanillajs
hand-written
PraxisJS
runtime 0.8.0
Solid
1.9.3
Svelte
5.42.1
Vue
3.5.39
Lit
3.2.0
Preact + signals
11
React hooks
19.2
uncompressed size
kB
11.3
1.00x
31.2
2.76x
11.5
1.02x
26.6
2.35x
64.3
5.69x
22.1
1.96x
23.3
2.06x
190.3
16.84x
compressed size
kB
2.5
1.00x
10.3
4.12x
4.5
1.80x
9.7
3.88x
23.3
9.32x
7.3
2.92x
8.3
3.32x
51.4
20.56x
first paint
ms
61.8
1.02x
81.5
1.35x
60.4
1.00x
86.8
1.44x
122.2
2.02x
70.7
1.17x
78.4
1.30x
278.5
4.61x

PraxisJS is not first on memory and size: it holds a little more memory than Solid, and its runtime (signals, decorators, components, keyed lists) ships more JavaScript than Solid's or Svelte's. First paint is ahead of Svelte, Vue and React, and behind Solid, Lit and Preact.

Script time

Script time is the JavaScript part of a benchmark: the framework's own work, without layout and paint.

Script time, median in ms
Script time (ms)PraxisJSSolidSvelteVueLitPreact + signalsReact hooksvanillajs
create 1k2.73.13.56.05.310.76.32.0
replace 1k5.06.07.19.78.419.212.63.9
update every 10th0.91.72.03.32.21.55.40.7
select row0.71.23.52.03.42.43.80.6
swap rows0.61.52.02.22.33.824.10.1
remove row0.50.60.93.01.21.41.70.5
create 10k25.526.929.446.945.198.6196.719.2
append 1k2.93.53.76.15.911.27.72.0
clear10.111.711.614.815.317.020.09.1

The runner starts the V8 CPU profiler and forces a full garbage collection right before each measured click. Code then runs colder than in a normal page, so script times are higher than what you would measure by hand. The conditions are the same for every framework.

Methodology

Date
October 16, 2026
Suite
js-framework-benchmark, keyed table
Browser
Chrome 155, visible window
Machine
Apple M3 Pro, macOS 27
Iterations
15 per benchmark, median reported
Run
PraxisJS 0.8 from npm
  • Each benchmark runs on a fresh page after warm-up runs. The suite slows the CPU down for some operations, the same way for every framework.
  • Every entry is the suite's own app for that framework, built for production. The PraxisJS entry is plain idiomatic code: a component with @State fields, rows as stores with an @State() label, For for the list and @Selector for the selected row.
  • All frameworks run in the same session on the same machine.

Why it is fast

The techniques behind these numbers, in plain language.

Writing fast PraxisJS

Practical rules for keeping your own app fast, and how to find slow parts.

Why it is fast

The techniques behind PraxisJS performance — fine-grained signals, compiled templates, keyed lists with minimal moves, delegated events, O(1) selection and V8-friendly objects.

On this page

BenchmarksHow to read thisResultsScript timeMethodology