Better immutability
Faster immutable updates

A JavaScript library for efficient immutable updates. Write plain mutations on a draft, and get the next immutable state.

Get Started GitHub0,000

$npm install mutative

Time for the same updates

Shorter is better

Mutative
1x
Immer
3.6x
Immer, auto-freeze
6.6x

Geometric means over 97 workloads, with the same settings for both libraries, and with each library's defaults, where only Immer freezes. Mutative was faster in 595 of 614 measured cases. See the benchmarks

Simple

Immutable updates without the spread

Mark the last item of a list as done and add a new one. With spread syntax, you copy each level of the state yourself; with Mutative, you change a draft.

Spread syntax

reducer.ts

const nextState = {
  ...state,
  list: [
    ...state.list.slice(0, 2),
    {
      ...state.list[2],
      done: true,
    },
    { text: 'Learn Mutative', done: true },
  ],
};

With Mutative

reducer.ts

import { create } from 'mutative';

const nextState = create(state, (draft) => {
  draft.list[2].done = true;
  draft.list.push({ text: 'Learn Mutative', done: true });
});

Fast

Faster than Immer, by design

Mutative copies each object at most once per update and moves array elements natively on its copy, instead of through a draft proxy.

3.6x

Faster than Immer with matched settings

Faster in 595 of 614 measured cases

6.6x

Faster than Immer with each library's defaults

Faster in 154 of 157 measured cases

Benchmarked workloads

Every result checked against hand-written reducers

Time per update, relative to Mutative

Shorter is better

Mutative
Immer
Immer, auto-freeze

200 RTK Query-style updates

100 pending and 100 resolved requests

1.1x faster

Mutative

Immer

1.1x

Immer, auto-freeze

4.8x

Update one field of a small object

A tiny plain object

1.4x faster

Mutative

Immer

1.4x

Immer, auto-freeze

1.9x

Add a number to a Set

A Set of 10,000 numbers

31x faster

Mutative

Immer

31x

Immer, auto-freeze

33x

Remove the first row

An array of 10,000 rows, with shift()

1,052x faster

Mutative

Immer

1,052x

Immer, auto-freeze

1,156x

Microseconds per update, shorter is better
Workload	Mutative	Immer	Immer, auto-freeze	Faster than Immer
200 RTK Query-style updates100 pending and 100 resolved requests	1,001 µs	1,133 µs	4,795 µs	1.1x
Update one field of a small objectA tiny plain object	0.51 µs	0.69 µs	0.95 µs	1.4x
Add a number to a SetA Set of 10,000 numbers	29.5 µs	904 µs	975 µs	31x
Remove the first rowAn array of 10,000 rows, with shift()	7.04 µs	7,408 µs	8,137 µs	1,052x

Medians of three runs on an Apple M1 Max with Node.js 24.16.0 and Immer 11.1.18, each library in a process of its own, without patches. Each bar is a library's time per update relative to Mutative's in its workload, and the table lists the times in microseconds. Mutative and the first Immer bar run without auto-freeze; the second Immer bar shows Immer's default. Hand-written reducers that copy only what they change are faster still in most workloads.

The factors of the cards are geometric means of Immer's time over Mutative's across all 97 workloads. See the benchmarks

In production