Benchmarks

@urcolor/core against the color libraries people actually reach for, plus the canvas grid samplers @urcolor/shared layers on top. Every number comes from the suites in packages/core/bench, regenerated with bun run --cwd packages/core bench:report.

Read the groups, not the totals

No library wins everywhere. These libraries trade spec coverage, perceptual accuracy and object model against each other, so a group urcolor loses says something about one of those choices rather than about the library. Where a comparison is not like for like, the note above the group says so.

A library missing from a group cannot express that operation at all. Nothing was dropped for being slow.

Each chart plots the fast end of its group. Bars past 10× the leader are named underneath and printed in the table, because one very slow bar flattens every other one to a pixel. Bar labels are in the unit named on the axis, rounded to three figures; the table under each chart has the rest.

Every library keeps the same color on every chart:

urcolorculorichroma-jscolorjs.iocolordtinycolor2@ctrl/tinycolor

Setup

Measured on Apple M1 at ~0.91 GHz, bun 1.3.0 (arm64-darwin), with mitata. Times are the mean per operation.

Library	Version
@urcolor/core	2.0.0
@urcolor/shared	1.0.0
culori	4.0.2
chroma-js	3.2.0
colorjs.io	0.7.1
colord	2.9.3
tinycolor2	1.6.0
@ctrl/tinycolor	4.2.0

# every suite
bun run --cwd packages/core bench

# one suite at a time
bun run --cwd packages/core bench parse convert

# regenerate this page
bun run --cwd packages/core bench:report

Each row runs against a rotating pool of eight colors. Against a single constant the JIT inlines the smaller operations and hoists them out of the measurement loop: an early draft of this suite had tinycolor2.toRgbString() "running" in 0.17 picoseconds. The rotation costs about a nanosecond, and every row pays it.

Two APIs, two costs

The same engine ships twice: a tree-shakeable functional API (parse, convert, mix) and the Color class on top of it. Groups listing both rows show what the wrapper costs, usually one allocation. Reach for the functions in per-pixel loops and the class everywhere else.

Parsing

CSS string in, the library's own representation out. Hex and the legacy functional notations are the common denominator. oklch(), lab() and color(display-p3 …) are CSS Color 4, which only urcolor, culori and colorjs.io parse.

parse: hex

Off the chart: colorjs.io at 25.07 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor parse (fn)	343.1 ns	fastest
culori parse	405.7 ns	1.18× slower
colord	411.4 ns	1.20× slower
urcolor Color.parse	471.9 ns	1.38× slower
chroma-js	1.69 µs	4.92× slower
tinycolor2	2.00 µs	5.82× slower
@ctrl/tinycolor	2.98 µs	8.69× slower
colorjs.io	25.07 µs	73.05× slower

parse: hex + alpha (#rrggbbaa)

Off the chart: colorjs.io at 12.45 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord	521.4 ns	fastest
urcolor	650.9 ns	1.25× slower
culori	952.5 ns	1.83× slower
@ctrl/tinycolor	1.57 µs	3.02× slower
tinycolor2	1.65 µs	3.16× slower
chroma-js	1.96 µs	3.75× slower
colorjs.io	12.45 µs	23.88× slower

parse: rgb() modern syntax

Off the chart: colorjs.io at 32.45 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord	1.19 µs	fastest
urcolor	1.80 µs	1.51× slower
culori	4.93 µs	4.16× slower
chroma-js	9.76 µs	8.23× slower
colorjs.io	32.45 µs	27.37× slower

parse: rgba() legacy syntax

Off the chart: colorjs.io at 17.93 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord	589.1 ns	fastest
@ctrl/tinycolor	881.1 ns	1.50× slower
culori	1.37 µs	2.32× slower
tinycolor2	1.58 µs	2.68× slower
urcolor	2.29 µs	3.89× slower
colorjs.io	17.93 µs	30.43× slower

parse: hsl()

Exact timings

Library	Time	Relative
tinycolor2	1.14 µs	fastest
urcolor	1.26 µs	1.11× slower
@ctrl/tinycolor	1.29 µs	1.13× slower
colord	1.35 µs	1.19× slower
culori	2.22 µs	1.95× slower
chroma-js	7.07 µs	6.22× slower
colorjs.io	8.31 µs	7.32× slower

parse: named color

Off the chart: colorjs.io at 4.97 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor	127.0 ns	fastest
colord	131.1 ns	1.03× slower
chroma-js	282.8 ns	2.23× slower
culori	397.2 ns	3.13× slower
tinycolor2	496.9 ns	3.91× slower
@ctrl/tinycolor	632.1 ns	4.98× slower
colorjs.io	4.97 µs	39.16× slower

parse: oklch() [CSS Color 4]

Off the chart: colorjs.io at 10.67 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor	676.0 ns	fastest
culori	1.29 µs	1.91× slower
colorjs.io	10.67 µs	15.79× slower

parse: lab() [CSS Color 4]

Exact timings

Library	Time	Relative
urcolor	851.1 ns	fastest
culori	1.39 µs	1.63× slower
colorjs.io	6.79 µs	7.97× slower

parse: color(display-p3 …) [CSS Color 4]

Off the chart: colorjs.io at 11.35 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor	796.1 ns	fastest
culori	1.42 µs	1.79× slower
colorjs.io	11.35 µs	14.25× slower

parse: invalid input (rejection)

Exact timings

Library	Time	Relative
urcolor tryParse	61.4 ns	fastest
colord .isValid()	156.1 ns	2.54× slower
@ctrl/tinycolor	355.8 ns	5.79× slower
culori parse	359.1 ns	5.85× slower
tinycolor2	369.9 ns	6.02× slower

Conversion

An already-parsed color moved into another space: transfer functions and matrices, with no parsing mixed in. Where a group converts out of a perceptual space, every operand is pre-converted, so each library pays for exactly one leg.

convert: sRGB → HSL

Off the chart: @ctrl/tinycolor at 342.7 ns, colorjs.io at 1.79 µs. The table below carries every row.

Exact timings

Library	Time	Relative
culori	21.4 ns	fastest
urcolor convert (fn)	43.0 ns	2.01× slower
colord	44.6 ns	2.08× slower
tinycolor2	71.0 ns	3.31× slower
urcolor .to()	71.4 ns	3.33× slower
chroma-js	94.3 ns	4.41× slower
@ctrl/tinycolor	342.7 ns	16.00× slower
colorjs.io	1.79 µs	83.66× slower

convert: sRGB → Oklch

Exact timings

Library	Time	Relative
culori	174.0 ns	fastest
urcolor convert (fn)	175.8 ns	1.01× slower
urcolor .to()	260.7 ns	1.50× slower
chroma-js	1.13 µs	6.47× slower
colorjs.io	1.32 µs	7.60× slower

convert: sRGB → Oklab

Off the chart: colorjs.io at 1.59 µs. The table below carries every row.

Exact timings

Library	Time	Relative
culori	117.5 ns	fastest
urcolor	149.3 ns	1.27× slower
chroma-js	1.17 µs	9.92× slower
colorjs.io	1.59 µs	13.57× slower

convert: sRGB → CIE Lab

Off the chart: colorjs.io at 1.38 µs. The table below carries every row.

Exact timings

Library	Time	Relative
culori	121.2 ns	fastest
colord	137.2 ns	1.13× slower
urcolor	182.6 ns	1.51× slower
chroma-js	220.6 ns	1.82× slower
colorjs.io	1.38 µs	11.41× slower

convert: sRGB → XYZ D65

Off the chart: colorjs.io at 1.22 µs. The table below carries every row.

Exact timings

Library	Time	Relative
culori	71.7 ns	fastest
urcolor	105.4 ns	1.47× slower
colorjs.io	1.22 µs	17.04× slower

convert: Oklch → sRGB

Exact timings

Library	Time	Relative
culori	145.9 ns	fastest
urcolor	237.3 ns	1.63× slower
colorjs.io	1.23 µs	8.44× slower
chroma-js	1.45 µs	9.92× slower

convert: sRGB → Display P3

Exact timings

Library	Time	Relative
culori	147.7 ns	fastest
urcolor	175.6 ns	1.19× slower
colorjs.io	1.21 µs	8.17× slower

convert: sRGB → Rec. 2020

Exact timings

Library	Time	Relative
culori	148.7 ns	fastest
urcolor	179.4 ns	1.21× slower
colorjs.io	1.31 µs	8.80× slower

convert: chain sRGB → Oklch → Lab → sRGB

Exact timings

Library	Time	Relative
culori	585.5 ns	fastest
urcolor	624.5 ns	1.07× slower
colorjs.io	4.10 µs	7.00× slower

gamut: map into sRGB

Off the chart: colorjs.io .toGamut() at 1.82 µs. The table below carries every row.

Exact timings

Library	Time	Relative
culori clampChroma	122.6 ns	fastest
urcolor .toGamut()	225.1 ns	1.84× slower
colorjs.io .toGamut()	1.82 µs	14.84× slower

gamut: inGamut check

Exact timings

Library	Time	Relative
culori	124.8 ns	fastest
urcolor	265.1 ns	2.12× slower
colorjs.io	863.7 ns	6.92× slower

Mixing and interpolation

One-shot blends, then the gradient shape: build an interpolator once and sample it repeatedly. colord and the tinycolors mix only in sRGB, which is cheaper and perceptually worse than an Oklab blend, so they appear in the sRGB group alone.

mix: sRGB, 50%

Off the chart: colorjs.io at 3.83 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor mix (fn)	137.8 ns	fastest
urcolor .mix()	204.4 ns	1.48× slower
chroma-js	422.4 ns	3.07× slower
colord	532.5 ns	3.86× slower
culori	757.3 ns	5.50× slower
@ctrl/tinycolor	1.11 µs	8.04× slower
tinycolor2	1.24 µs	9.01× slower
colorjs.io	3.83 µs	27.81× slower

mix: Oklab, 50%

Exact timings

Library	Time	Relative
urcolor mix (fn)	419.1 ns	fastest
urcolor .mix()	446.4 ns	1.06× slower
culori	835.9 ns	1.99× slower
colorjs.io	3.24 µs	7.74× slower
chroma-js	3.80 µs	9.06× slower

mix: Oklch (shorter hue arc), 50%

Off the chart: colorjs.io at 8.25 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor	489.0 ns	fastest
culori	897.3 ns	1.84× slower
chroma-js	4.56 µs	9.33× slower
colorjs.io	8.25 µs	16.87× slower

mix: CIE Lab, 50%

Off the chart: colorjs.io at 3.86 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor	367.6 ns	fastest
culori	902.2 ns	2.45× slower
chroma-js	1.21 µs	3.30× slower
colorjs.io	3.86 µs	10.51× slower

gradient: 64 Oklab samples (interpolator reused)

Off the chart: colorjs.io at 36.10 µs, chroma-js at 235.94 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor	2.70 µs	fastest
culori	6.43 µs	2.38× slower
colorjs.io	36.10 µs	13.39× slower
chroma-js	235.94 µs	87.53× slower

gradient: 64 Oklab samples (mix per sample)

Off the chart: chroma-js at 225.04 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor	20.77 µs	fastest
culori	52.03 µs	2.50× slower
colorjs.io	197.01 µs	9.48× slower
chroma-js	225.04 µs	10.83× slower

palette: 11 stops from two anchors

Off the chart: colorjs.io at 39.14 µs, chroma-js at 39.42 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor	3.31 µs	fastest
culori	3.35 µs	1.01× slower
colorjs.io	39.14 µs	11.81× slower
chroma-js	39.42 µs	11.90× slower

Manipulation

The libraries disagree about where an adjustment happens. urcolor and chroma-js lighten in a perceptual space and pay for two conversions on top of the arithmetic. colord and the tinycolors nudge HSL lightness directly. Same verb, different job.

manipulate: lighten (perceptual: urcolor/chroma-js; HSL: rest)

Off the chart: tinycolor2 at 1.72 µs, @ctrl/tinycolor at 1.83 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord .lighten()	131.9 ns	fastest
urcolor lighten (fn)	342.0 ns	2.59× slower
urcolor .lighten()	457.1 ns	3.47× slower
chroma-js .brighten()	493.8 ns	3.74× slower
tinycolor2	1.72 µs	13.06× slower
@ctrl/tinycolor	1.83 µs	13.87× slower

manipulate: darken

Off the chart: tinycolor2 at 2.09 µs, @ctrl/tinycolor at 2.25 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord	128.1 ns	fastest
urcolor	356.3 ns	2.78× slower
chroma-js	586.9 ns	4.58× slower
tinycolor2	2.09 µs	16.28× slower
@ctrl/tinycolor	2.25 µs	17.53× slower

manipulate: saturate

Off the chart: tinycolor2 at 1.51 µs, @ctrl/tinycolor at 1.86 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord	119.3 ns	fastest
urcolor	348.0 ns	2.92× slower
chroma-js	644.6 ns	5.41× slower
tinycolor2	1.51 µs	12.69× slower
@ctrl/tinycolor	1.86 µs	15.62× slower

manipulate: rotate hue 60°

Off the chart: tinycolor2 at 1.80 µs, @ctrl/tinycolor at 2.08 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord	146.5 ns	fastest
urcolor	356.3 ns	2.43× slower
tinycolor2	1.80 µs	12.30× slower
@ctrl/tinycolor	2.08 µs	14.23× slower

channel: set lightness in Oklch

Exact timings

Library	Time	Relative
urcolor .with()	210.3 ns	fastest
colorjs.io .set()	2.13 µs	10.14× slower

channel: read one channel

Off the chart: chroma-js at 96.0 ns, colorjs.io at 666.5 ns. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor .get()	7.3 ns	fastest
colord	18.6 ns	2.54× slower
chroma-js	96.0 ns	13.08× slower
colorjs.io	666.5 ns	90.84× slower

alpha: set opacity

Off the chart: @ctrl/tinycolor at 712.9 ns, tinycolor2 at 726.8 ns. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor .withAlpha()	27.3 ns	fastest
colord .alpha()	73.7 ns	2.70× slower
chroma-js .alpha()	253.2 ns	9.29× slower
@ctrl/tinycolor	712.9 ns	26.14× slower
tinycolor2	726.8 ns	26.65× slower

Difference and contrast

ΔE2000 is the expensive one: trigonometry plus a rotation term. ΔE76 is here so the cost of the 2000 formula is visible rather than implied. WCAG contrast is a handful of multiplications, so that group mostly measures each library's object-model overhead.

deltaE: CIEDE2000

Exact timings

Library	Time	Relative
culori	384.1 ns	fastest
urcolor deltaE (fn)	412.3 ns	1.07× slower
urcolor .deltaE()	432.6 ns	1.13× slower
colord .delta()	558.9 ns	1.46× slower
chroma-js	654.0 ns	1.70× slower
colorjs.io	1.71 µs	4.44× slower

deltaE: CIE76 (Euclidean Lab)

Exact timings

Library	Time	Relative
urcolor	290.4 ns	fastest
culori	353.2 ns	1.22× slower
colorjs.io	1.23 µs	4.24× slower

deltaE: ΔEOK (Oklab Euclidean)

Exact timings

Library	Time	Relative
urcolor	260.5 ns	fastest
culori	276.3 ns	1.06× slower
colorjs.io	1.20 µs	4.62× slower

contrast: WCAG 2.1 ratio

Off the chart: colorjs.io at 2.97 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor contrast (fn)	140.2 ns	fastest
culori	151.6 ns	1.08× slower
urcolor .contrast()	156.2 ns	1.11× slower
colord	262.1 ns	1.87× slower
chroma-js	302.6 ns	2.16× slower
tinycolor2	1.12 µs	8.01× slower
@ctrl/tinycolor	1.25 µs	8.89× slower
colorjs.io	2.97 µs	21.21× slower

contrast: APCA

Exact timings

Library	Time	Relative
urcolor	196.2 ns	fastest
colorjs.io	1.70 µs	8.65× slower

relative luminance

Exact timings

Library	Time	Relative
colord	55.4 ns	fastest
culori	69.3 ns	1.25× slower
chroma-js	93.4 ns	1.69× slower
urcolor	97.9 ns	1.77× slower

equality: same color, different space

Exact timings

Library	Time	Relative
chroma-js hex compare	394.2 ns	fastest
colord .isEqual()	433.3 ns	1.10× slower
urcolor .equals()	445.4 ns	1.13× slower

Serialization

Color object to CSS string, the last step of every render path. It runs as often as parsing does.

serialize: → hex

Off the chart: colorjs.io at 619.0 ns. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor serialize (fn)	36.6 ns	fastest
urcolor .toString()	45.6 ns	1.25× slower
culori formatHex	57.4 ns	1.57× slower
colord .toHex()	84.3 ns	2.30× slower
tinycolor2	113.8 ns	3.11× slower
@ctrl/tinycolor	118.8 ns	3.25× slower
chroma-js .hex()	144.9 ns	3.96× slower
colorjs.io	619.0 ns	16.92× slower

serialize: → rgb()

Off the chart: chroma-js at 246.5 ns, colorjs.io at 791.6 ns. The table below carries every row.

Exact timings

Library	Time	Relative
tinycolor2	11.2 ns	fastest
@ctrl/tinycolor	12.8 ns	1.14× slower
culori	34.7 ns	3.09× slower
colord	39.0 ns	3.48× slower
urcolor	50.1 ns	4.47× slower
chroma-js	246.5 ns	21.96× slower
colorjs.io	791.6 ns	70.54× slower

serialize: → hsl()

Off the chart: colorjs.io at 2.01 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord	55.3 ns	fastest
tinycolor2	79.7 ns	1.44× slower
urcolor	127.3 ns	2.30× slower
culori	136.9 ns	2.47× slower
@ctrl/tinycolor	323.9 ns	5.86× slower
colorjs.io	2.01 µs	36.40× slower

serialize: → oklch()

Off the chart: colorjs.io at 1.99 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor	59.3 ns	fastest
culori	61.8 ns	1.04× slower
colorjs.io	1.99 µs	33.49× slower

serialize: sRGB color → oklch() string (convert + print)

Exact timings

Library	Time	Relative
culori	217.5 ns	fastest
urcolor	280.9 ns	1.29× slower
colorjs.io	1.96 µs	9.00× slower

Canvas gradient rendering

The CPU path behind a picker surface: a slider track, an SV plane, a hue wheel. Each call returns an Uint8ClampedArray of RGBA bytes ready for putImageData, so these are milliseconds per frame. No other library ships a grid sampler, so the comparison rows hand-roll the same loop with that library's interpolator. urcolor's WebGL entry points are not measured here: they need a live canvas and a GPU context, and their cost is one uniform upload and one draw call whatever the surface size.

canvas: 512px linear track, Oklab → RGBA bytes

Off the chart: chroma-js (hand-rolled) at 1.89 ms. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor interpolateStops	93.75 µs	fastest
culori (hand-rolled)	108.43 µs	1.16× slower
colorjs.io (hand-rolled)	831.88 µs	8.87× slower
chroma-js (hand-rolled)	1.89 ms	20.19× slower

canvas: 128×128 bilinear plane, Oklab → RGBA bytes

Off the chart: colorjs.io (hand-rolled) at 178.23 ms. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor sampleBilinearGrid	4.18 ms	fastest
culori (hand-rolled)	31.07 ms	7.43× slower
colorjs.io (hand-rolled)	178.23 ms	42.65× slower

canvas: 128×128 HSV S/V plane → RGBA bytes

Off the chart: colorjs.io (hand-rolled) at 30.45 ms. The table below carries every row.

Exact timings

Library	Time	Relative
culori (hand-rolled)	1.11 ms	fastest
urcolor sampleChannelGrid	1.81 ms	1.63× slower
colorjs.io (hand-rolled)	30.45 ms	27.54× slower

canvas: 128×128 Oklch polar wheel → RGBA bytes

Exact timings

Library	Time	Relative
culori (hand-rolled)	2.85 ms	fastest
urcolor samplePolarGrid	3.56 ms	1.25× slower

canvas: 128×128 conic hue ring → RGBA bytes

Exact timings

Library	Time	Relative
culori (hand-rolled)	2.93 ms	fastest
urcolor sampleConicRing	3.78 ms	1.29× slower

canvas: per-pixel cost, 1 000 Oklch → sRGB pixels

Off the chart: colorjs.io (hand-rolled) at 1.99 ms. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor sampleConicRing	135.22 µs	fastest
culori (hand-rolled)	163.91 µs	1.21× slower
chroma-js (hand-rolled)	1.31 ms	9.68× slower
colorjs.io (hand-rolled)	1.99 ms	14.73× slower

End-to-end pipelines

String in, string out. The groups above isolate one cost each; these measure the round trip an application actually pays, which is where a library's object model shows up.

pipeline: hex → lighten → hex

Off the chart: colorjs.io at 30.32 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord	409.3 ns	fastest
urcolor fn	510.4 ns	1.25× slower
culori (hand-rolled)	520.8 ns	1.27× slower
urcolor Color	589.0 ns	1.44× slower
chroma-js	1.13 µs	2.77× slower
tinycolor2	1.51 µs	3.69× slower
@ctrl/tinycolor	2.11 µs	5.17× slower
colorjs.io	30.32 µs	74.08× slower

pipeline: hex → oklch → hex (round trip)

Off the chart: colorjs.io at 8.04 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor fn	488.0 ns	fastest
culori	495.4 ns	1.02× slower
urcolor Color	607.2 ns	1.24× slower
chroma-js	3.11 µs	6.37× slower
colorjs.io	8.04 µs	16.47× slower

pipeline: two hex strings → mixed hex (Oklab)

Off the chart: colorjs.io at 15.89 µs. The table below carries every row.

Exact timings

Library	Time	Relative
urcolor fn	748.5 ns	fastest
urcolor Color	859.4 ns	1.15× slower
culori	1.39 µs	1.85× slower
chroma-js	5.71 µs	7.63× slower
colorjs.io	15.89 µs	21.22× slower

pipeline: two hex strings → ΔE2000

Off the chart: colorjs.io at 9.60 µs. The table below carries every row.

Exact timings

Library	Time	Relative
culori	676.0 ns	fastest
colord	700.7 ns	1.04× slower
urcolor fn	717.2 ns	1.06× slower
urcolor Color	803.2 ns	1.19× slower
chroma-js	1.54 µs	2.28× slower
colorjs.io	9.60 µs	14.20× slower

pipeline: two hex strings → WCAG contrast

Off the chart: colorjs.io at 11.09 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord	408.3 ns	fastest
urcolor fn	478.0 ns	1.17× slower
culori	518.1 ns	1.27× slower
urcolor Color	535.0 ns	1.31× slower
tinycolor2	964.4 ns	2.36× slower
chroma-js	1.19 µs	2.90× slower
colorjs.io	11.09 µs	27.17× slower

batch: 10 swatches → darken → hex

Off the chart: colorjs.io at 188.20 µs. The table below carries every row.

Exact timings

Library	Time	Relative
colord	3.65 µs	fastest
urcolor fn	5.18 µs	1.42× slower
culori (hand-rolled)	5.22 µs	1.43× slower
urcolor Color	6.05 µs	1.66× slower
chroma-js	13.25 µs	3.63× slower
tinycolor2	16.88 µs	4.62× slower
colorjs.io	188.20 µs	51.57× slower

Reproducing

A benchmark is worth what the machine running it is worth. To check these numbers yourself:

git clone https://github.com/ur-color/urcolor
cd urcolor && bun install
bun run --cwd packages/core bench

Close everything else first, because background load lands squarely in the mean. If your ordering differs from these tables and you can reproduce it, open an issue. A benchmark that holds on one laptop only is not a benchmark.