Kanon Performance Benchmarks

Real-world validation benchmarks. Data auto-generated on Mar 14, 2026.

TL;DR

Kanon JIT dominates on discriminated unions and complex schemas. TypeBox/AJV win on simple validations. For typical API/form validation, Kanon is 2-3x faster than Zod/Valibot.

Operations per second. Higher is better.

---

Methodology

Each benchmark:

1. Creates a pool of 10,000 test objects
2. Runs validation in a tight loop
3. Measures operations per second with statistical analysis
4. Reports min, max, mean, percentiles (p75, p99, p995, p999)
5. Includes relative margin of error (rme) for confidence

The "fastest" label is only awarded when the winner is ≥10% faster than the runner-up. Otherwise, it's considered a statistical tie.

Does Performance Matter?

Not all validations matter equally for performance. Validating an API response called 1000x/sec matters more than a one-time config check. We assign weights based on real-world usage patterns:

Perf matters?	Weight	Description	Examples
CRITICAL A lot	5 pts	Hot path functions called in tight loops	map, filter, reduce, groupBy, chunk
HIGH Yes	3 pts	Frequently used utilities, but not in tight loops	get, set, pick, omit, cloneDeep
MEDIUM A bit	1 pt	Occasional utilities	Type guards (isArray, isString), string formatting
LOW No	0.5 pts	Setup-only functions where runtime perf is irrelevant	debounce, throttle, once, memoize

This scoring gives a more realistic picture of which library will actually make your app faster.

---

Libraries Tested

Library	Version	Description
@sinclair/typebox	0.34.48	JSON Schema with TypeScript inference + JIT
ajv	8.18.0	JSON Schema validator with JIT compilation
effect	3.19.19	Full-featured TypeScript library with schema validation
fastest-validator	1.19.1	High-performance validator with JIT
kanon	2.3.0	Pithos validation module (schema-first + JIT)
superstruct	2.0.2	Composable validation with custom types
valibot	1.2.0	Modular schema validation, tree-shakable
zod	4.3.6	Schema-first validation with TypeScript inference

Benchmarks run on Mar 14, 2026

---

Benchmark Results

Groups:

Schema-firstCompiled / JIT

Libraries:

@kanon/V3.0ZodValibotSuperstructEffect@kanon/JITAJVTypeBoxFast-Validator

Test	@kanon/V3.0	@kanon/JIT	AJV	Effect	Fast-Validator	Superstruct	TypeBox	Valibot	Zod
Critical
Login Form Validation CRITICAL | 5 pts	7.30M ops/s (2.22x)	10.75M ops/s (1.51x)	8.85M ops/s (1.83x)	2.69M ops/s (6.03x)	10.30M ops/s (1.57x)	606.72K ops/s (26.74x)	16.22M ops/s fastest	3.35M ops/s (4.85x)	4.84M ops/s (3.35x)
Payment Form (conditional validation) CRITICAL | 5 pts	5.54M ops/s (3.54x)	19.57M ops/s fastest	16.07M ops/s (1.22x)	861.76K ops/s (22.71x)	12.53M ops/s (1.56x)	298.76K ops/s (65.51x)	1.74M ops/s (11.24x)	1.68M ops/s (11.63x)	2.62M ops/s (7.48x)
User Registration (with password confirm) CRITICAL | 5 pts	5.04M ops/s (1.83x)	6.37M ops/s (1.45x)	6.95M ops/s (1.33x)	1.04M ops/s (8.82x)	6.91M ops/s (1.33x)	302.28K ops/s (30.47x)	9.21M ops/s fastest	1.68M ops/s (5.48x)	1.96M ops/s (4.71x)
High
API Response (discriminated union) HIGH | 3 pts	8.54M ops/s (3.06x)	26.14M ops/s fastest	18.93M ops/s (1.38x)	2.88M ops/s (9.07x)	15.24M ops/s (1.72x)	374.49K ops/s (69.8x)	3.38M ops/s (7.74x)	2.40M ops/s (10.88x)	8.23M ops/s (3.17x)
Blog Post with Comments HIGH | 3 pts	2.53M ops/s (2.64x)	6.70M ops/s fastest	5.09M ops/s (1.32x)	570.87K ops/s (11.73x)	3.64M ops/s (1.84x)	110.87K ops/s (60.39x)	1.49M ops/s (4.48x)	794.08K ops/s (8.43x)	1.48M ops/s (4.52x)
E-commerce Product HIGH | 3 pts	4.93M ops/s (2.92x)	14.40M ops/s fastest	12.74M ops/s (1.13x)	803.89K ops/s (17.92x)	12.47M ops/s (1.15x)	198.25K ops/s (72.65x)	1.13M ops/s (12.72x)	1.39M ops/s (10.32x)	2.58M ops/s (5.57x)
Event Booking HIGH | 3 pts	1.65M ops/s (5.57x)	9.16M ops/s fastest	4.03M ops/s (2.27x)	379.42K ops/s (24.15x)	2.91M ops/s (3.15x)	120.28K ops/s (76.18x)	2.79M ops/s (3.29x)	686.53K ops/s (13.35x)	839.21K ops/s (10.92x)
Invalid Login (error handling) HIGH | 3 pts	5.76M ops/s (1.63x)	9.26M ops/s (1.01x)	9.38M ops/s fastest	265.91K ops/s (35.28x)	1.98M ops/s (4.74x)	189.99K ops/s (49.37x)	7.98M ops/s (1.18x)	2.16M ops/s (4.34x)	149.75K ops/s (62.64x)
Medium
Search Params (with coercion) MEDIUM | 1 pts	5.05M ops/s (4.64x)	23.37M ops/s (1x)	23.43M ops/s fastest	1.36M ops/s (17.27x)	18.07M ops/s (1.3x)	190.53K ops/s (122.97x)	2.01M ops/s (11.68x)	1.77M ops/s (13.24x)	4.19M ops/s (5.59x)
User Profile Update (optional fields) MEDIUM | 1 pts	6.02M ops/s (3.08x)	18.58M ops/s fastest	7.88M ops/s (2.36x)	2.72M ops/s (6.84x)	11.62M ops/s (1.6x)	453.75K ops/s (40.95x)	5.32M ops/s (3.49x)	2.34M ops/s (7.95x)	2.08M ops/s (8.94x)

Data generated on Mar 14, 2026 • Vitest bench

📊 Performance Summary

@kanon/JIT

6 wins

TypeBox

2 wins

AJV

2 wins

@kanon/V3.0

0 wins

Zod

0 wins

Valibot

0 wins

Superstruct

0 wins

Effect

0 wins

Fast-Validator

0 wins

📈 Weighted Summary

🏆 @kanon/JIT dominates real-world scenarios!

@kanon/JIT

59%(24 pts)

TypeBox

24%(10 pts)

AJV

17%(7 pts)

@kanon/V3.0

0%(0 pts)

Zod

0%(0 pts)

Valibot

0%(0 pts)

Superstruct

0%(0 pts)

Effect

0%(0 pts)

Fast-Validator

0%(0 pts)

On critical + high priority tasks, @kanon/JIT is 2.4x ahead.

Detailed Statistics

For the skeptics who want to see the raw numbers:

Login Form ValidationCRITICAL

Library	ops/s	min	max	mean	p75	p99	p995	p999	rme	samples
TypeBox 🏆	16.22M	0.0000	0.0160	0.0001	0.0001	0.0001	0.0001	0.0002	±0.03%	8,110,826
@kanon/JIT	10.75M	0.0000	0.0251	0.0001	0.0001	0.0001	0.0001	0.0002	±0.02%	5,375,293
Fast-Validator	10.30M	0.0000	0.0193	0.0001	0.0001	0.0001	0.0001	0.0002	±0.03%	5,151,905
AJV	8.85M	0.0000	0.3477	0.0001	0.0001	0.0001	0.0002	0.0002	±0.14%	4,426,595
@kanon/V3.0	7.30M	0.0000	0.0334	0.0001	0.0002	0.0002	0.0002	0.0003	±0.04%	3,649,270
Zod	4.84M	0.0001	0.3160	0.0002	0.0002	0.0003	0.0003	0.0005	±0.28%	2,418,545
Valibot	3.35M	0.0002	0.3557	0.0003	0.0003	0.0004	0.0004	0.0005	±0.33%	1,673,937
Effect	2.69M	0.0002	0.3195	0.0004	0.0004	0.0005	0.0005	0.0007	±0.44%	1,345,564
Superstruct	606.72K	0.0014	0.3060	0.0016	0.0016	0.0021	0.0022	0.0028	±0.61%	303,362

User Registration (with password confirm)CRITICAL

Library	ops/s	min	max	mean	p75	p99	p995	p999	rme	samples
TypeBox 🏆	9.21M	0.0000	0.3371	0.0001	0.0001	0.0001	0.0002	0.0003	±0.48%	4,605,625
AJV	6.95M	0.0000	0.2434	0.0001	0.0002	0.0002	0.0002	0.0003	±0.10%	3,473,459
Fast-Validator	6.91M	0.0000	0.7337	0.0001	0.0002	0.0002	0.0002	0.0003	±0.53%	3,456,964
@kanon/JIT	6.37M	0.0000	0.0193	0.0002	0.0002	0.0002	0.0002	0.0003	±0.03%	3,183,797
@kanon/V3.0	5.04M	0.0001	0.0120	0.0002	0.0002	0.0003	0.0003	0.0003	±0.02%	2,519,971
Zod	1.96M	0.0004	0.5719	0.0005	0.0005	0.0007	0.0007	0.0010	±0.71%	978,593
Valibot	1.68M	0.0005	0.4611	0.0006	0.0006	0.0007	0.0007	0.0008	±0.45%	840,156
Effect	1.04M	0.0008	0.3644	0.0010	0.0010	0.0012	0.0012	0.0014	±0.40%	522,297
Superstruct	302.28K	0.0029	0.4256	0.0033	0.0033	0.0039	0.0041	0.0056	±0.60%	151,139

API Response (discriminated union)HIGH

Library	ops/s	min	max	mean	p75	p99	p995	p999	rme	samples
@kanon/JIT 🏆	26.14M	0.0000	0.0310	0.0000	0.0000	0.0000	0.0001	0.0001	±0.03%	13,069,900
AJV	18.93M	0.0000	0.4032	0.0001	0.0000	0.0001	0.0001	0.0002	±0.43%	9,466,101
Fast-Validator	15.24M	0.0000	0.9756	0.0001	0.0001	0.0001	0.0001	0.0002	±0.38%	7,620,768
@kanon/V3.0	8.54M	0.0000	0.0126	0.0001	0.0001	0.0001	0.0002	0.0002	±0.02%	4,271,265
Zod	8.23M	0.0000	0.3493	0.0001	0.0001	0.0002	0.0002	0.0003	±0.38%	4,117,143
TypeBox	3.38M	0.0002	0.3417	0.0003	0.0003	0.0004	0.0004	0.0005	±0.34%	1,688,336
Effect	2.88M	0.0002	0.3228	0.0003	0.0003	0.0005	0.0005	0.0007	±0.49%	1,440,551
Valibot	2.40M	0.0003	0.3029	0.0004	0.0005	0.0005	0.0006	0.0006	±0.16%	1,201,783
Superstruct	374.49K	0.0020	0.3914	0.0027	0.0034	0.0039	0.0040	0.0050	±0.55%	187,246

E-commerce ProductHIGH

Library	ops/s	min	max	mean	p75	p99	p995	p999	rme	samples
@kanon/JIT 🏆	14.40M	0.0000	0.0372	0.0001	0.0001	0.0001	0.0001	0.0002	±0.03%	7,201,078
AJV	12.74M	0.0000	0.0109	0.0001	0.0001	0.0001	0.0001	0.0002	±0.02%	6,371,751
Fast-Validator	12.47M	0.0000	0.4601	0.0001	0.0001	0.0001	0.0001	0.0002	±0.40%	6,237,172
@kanon/V3.0	4.93M	0.0001	0.0544	0.0002	0.0002	0.0003	0.0003	0.0004	±0.04%	2,462,980
Zod	2.58M	0.0003	0.4524	0.0004	0.0004	0.0005	0.0006	0.0007	±0.50%	1,292,474
Valibot	1.39M	0.0005	0.3606	0.0007	0.0008	0.0009	0.0010	0.0011	±0.40%	697,498
TypeBox	1.13M	0.0007	0.4893	0.0009	0.0010	0.0011	0.0011	0.0013	±0.19%	566,269
Effect	803.89K	0.0010	0.2753	0.0012	0.0013	0.0015	0.0016	0.0018	±0.50%	401,946
Superstruct	198.25K	0.0043	0.3490	0.0050	0.0053	0.0060	0.0063	0.0091	±0.51%	99,123

Blog Post with CommentsHIGH

Library	ops/s	min	max	mean	p75	p99	p995	p999	rme	samples
@kanon/JIT 🏆	6.70M	0.0000	0.0270	0.0001	0.0002	0.0002	0.0002	0.0003	±0.05%	3,347,674
AJV	5.09M	0.0001	0.4245	0.0002	0.0002	0.0003	0.0003	0.0003	±0.17%	2,545,400
Fast-Validator	3.64M	0.0001	0.5038	0.0003	0.0003	0.0004	0.0005	0.0006	±0.20%	1,822,243
@kanon/V3.0	2.53M	0.0002	0.0180	0.0004	0.0005	0.0005	0.0006	0.0007	±0.04%	1,267,014
TypeBox	1.49M	0.0005	0.4178	0.0007	0.0007	0.0008	0.0008	0.0010	±0.32%	746,601
Zod	1.48M	0.0005	0.4118	0.0007	0.0007	0.0010	0.0010	0.0012	±0.47%	740,333
Valibot	794.08K	0.0007	0.3297	0.0013	0.0015	0.0018	0.0019	0.0022	±0.36%	397,041
Effect	570.87K	0.0012	0.3074	0.0018	0.0020	0.0024	0.0025	0.0029	±0.45%	285,436
Superstruct	110.87K	0.0053	0.2614	0.0090	0.0105	0.0136	0.0160	0.0328	±0.50%	55,436

Search Params (with coercion)MEDIUM

Library	ops/s	min	max	mean	p75	p99	p995	p999	rme	samples
AJV 🏆	23.43M	0.0000	0.5813	0.0000	0.0000	0.0001	0.0001	0.0001	±0.30%	11,715,355
@kanon/JIT	23.37M	0.0000	0.0342	0.0000	0.0000	0.0001	0.0001	0.0001	±0.03%	11,685,828
Fast-Validator	18.07M	0.0000	0.0195	0.0001	0.0001	0.0001	0.0001	0.0001	±0.03%	9,033,883
@kanon/V3.0	5.05M	0.0001	0.3721	0.0002	0.0002	0.0002	0.0003	0.0004	±0.45%	2,525,489
Zod	4.19M	0.0001	0.5748	0.0002	0.0002	0.0003	0.0004	0.0005	±0.54%	2,097,472
TypeBox	2.01M	0.0004	0.4598	0.0005	0.0005	0.0006	0.0006	0.0007	±0.28%	1,002,752
Valibot	1.77M	0.0005	0.2843	0.0006	0.0006	0.0007	0.0007	0.0009	±0.23%	884,974
Effect	1.36M	0.0006	0.3948	0.0007	0.0008	0.0009	0.0010	0.0011	±0.41%	678,193
Superstruct	190.53K	0.0048	2.7560	0.0052	0.0052	0.0079	0.0082	0.0096	±1.31%	95,268

User Profile Update (optional fields)MEDIUM

Library	ops/s	min	max	mean	p75	p99	p995	p999	rme	samples
@kanon/JIT 🏆	18.58M	0.0000	0.0374	0.0001	0.0001	0.0001	0.0001	0.0002	±0.05%	9,289,664
Fast-Validator	11.62M	0.0000	0.0276	0.0001	0.0001	0.0001	0.0001	0.0002	±0.03%	5,809,017
AJV	7.88M	0.0000	0.3157	0.0001	0.0002	0.0003	0.0003	0.0003	±0.14%	3,941,102
@kanon/V3.0	6.02M	0.0000	0.0182	0.0002	0.0002	0.0003	0.0003	0.0004	±0.04%	3,011,460
TypeBox	5.32M	0.0000	0.4100	0.0002	0.0002	0.0003	0.0003	0.0004	±0.43%	2,660,166
Effect	2.72M	0.0000	0.3086	0.0004	0.0005	0.0009	0.0009	0.0010	±0.43%	1,358,322
Valibot	2.34M	0.0002	0.3663	0.0004	0.0006	0.0008	0.0008	0.0009	±0.31%	1,168,216
Zod	2.08M	0.0002	18.9033	0.0005	0.0005	0.0007	0.0008	0.0009	±10.28%	1,038,614
Superstruct	453.75K	0.0015	0.2033	0.0022	0.0025	0.0032	0.0033	0.0040	±0.38%	226,873

Payment Form (conditional validation)CRITICAL

Library	ops/s	min	max	mean	p75	p99	p995	p999	rme	samples
@kanon/JIT 🏆	19.57M	0.0000	0.0114	0.0001	0.0000	0.0001	0.0001	0.0001	±0.03%	9,786,506
AJV	16.07M	0.0000	0.5031	0.0001	0.0001	0.0001	0.0001	0.0002	±0.20%	8,036,981
Fast-Validator	12.53M	0.0000	0.5377	0.0001	0.0001	0.0001	0.0001	0.0002	±0.53%	6,267,041
@kanon/V3.0	5.54M	0.0000	1.1010	0.0002	0.0002	0.0003	0.0003	0.0004	±0.43%	2,768,177
Zod	2.62M	0.0002	0.4883	0.0004	0.0005	0.0006	0.0006	0.0008	±0.63%	1,308,935
TypeBox	1.74M	0.0005	0.4601	0.0006	0.0006	0.0007	0.0008	0.0009	±0.30%	870,524
Valibot	1.68M	0.0005	0.0172	0.0006	0.0006	0.0007	0.0008	0.0008	±0.03%	841,630
Effect	861.76K	0.0009	0.2639	0.0012	0.0012	0.0014	0.0015	0.0017	±0.17%	430,879
Superstruct	298.76K	0.0029	0.2312	0.0033	0.0034	0.0039	0.0041	0.0054	±0.46%	149,382

Event BookingHIGH

Library	ops/s	min	max	mean	p75	p99	p995	p999	rme	samples
@kanon/JIT 🏆	9.16M	0.0000	0.1130	0.0001	0.0001	0.0002	0.0002	0.0002	±0.06%	4,581,214
AJV	4.03M	0.0000	0.3578	0.0002	0.0003	0.0004	0.0004	0.0005	±0.15%	2,017,384
Fast-Validator	2.91M	0.0001	0.3994	0.0003	0.0005	0.0006	0.0007	0.0008	±0.48%	1,452,835
TypeBox	2.79M	0.0002	0.4330	0.0004	0.0004	0.0005	0.0005	0.0007	±0.36%	1,394,009
@kanon/V3.0	1.65M	0.0002	0.0481	0.0006	0.0008	0.0010	0.0011	0.0013	±0.08%	822,949
Zod	839.21K	0.0005	0.4272	0.0012	0.0015	0.0019	0.0020	0.0023	±0.42%	419,605
Valibot	686.53K	0.0007	0.3650	0.0015	0.0018	0.0022	0.0023	0.0027	±0.36%	343,264
Effect	379.42K	0.0011	0.4886	0.0026	0.0033	0.0045	0.0049	0.0074	±0.63%	189,713
Superstruct	120.28K	0.0041	0.3946	0.0083	0.0102	0.0125	0.0131	0.0197	±0.55%	60,141

Invalid Login (error handling)HIGH

Library	ops/s	min	max	mean	p75	p99	p995	p999	rme	samples
AJV 🏆	9.38M	0.0000	0.0644	0.0001	0.0001	0.0002	0.0002	0.0002	±0.08%	4,690,265
@kanon/JIT	9.26M	0.0000	0.0189	0.0001	0.0001	0.0002	0.0002	0.0002	±0.03%	4,627,743
TypeBox	7.98M	0.0000	0.3721	0.0001	0.0001	0.0002	0.0002	0.0003	±0.46%	3,988,627
@kanon/V3.0	5.76M	0.0001	0.5792	0.0002	0.0002	0.0002	0.0003	0.0003	±0.23%	2,878,782
Valibot	2.16M	0.0002	0.4915	0.0005	0.0005	0.0007	0.0008	0.0009	±0.62%	1,080,612
Fast-Validator	1.98M	0.0000	0.3687	0.0005	0.0007	0.0012	0.0013	0.0015	±0.42%	989,654
Effect	265.91K	0.0004	0.4494	0.0038	0.0054	0.0067	0.0070	0.0076	±0.42%	132,956
Superstruct	189.99K	0.0018	20.0408	0.0053	0.0051	0.0075	0.0085	0.0172	±10.70%	94,996
Zod	149.75K	0.0002	9.1670	0.0067	0.0090	0.0118	0.0127	0.0190	±6.55%	74,873

---

Why Kanon JIT is Fast

1. JIT Compilation: Schemas are compiled to optimized JavaScript functions at runtime
2. No runtime type checks: Types are validated at compile-time by TypeScript
3. Minimal abstraction: Direct validation logic, no class hierarchies
4. Discriminated unions: O(1) lookup instead of O(n) trial-and-error

---

A Note on Coercion Performance

On basic coercion operations, Kanon and Zod are roughly equivalent: each wins some benchmarks. This is despite a deliberate architectural difference.

Zod mutates the input directly:

// Zod's approach (simplified)
if (def.coerce) {
  payload.value = new Date(payload.value); // mutation!
}

Kanon returns a new value:

// Kanon's approach
if (value instanceof Date) return true;
return { coerced: new Date(value) }; // pure, no mutation

We chose immutability over raw speed because:

• Predictability: Functions don't modify their arguments
• Debugging: Easier to trace where values come from
• Composition: Pure functions compose better
• Safety: No unexpected side-effects

The performance difference is negligible in practice (~0.00001ms per validation), and Kanon still dominates on coercion with constraints (36x faster than Zod on date.max()), which is the common real-world use case.

---

Reproduce These Results

Want to verify these results? See how to reproduce our data.

---

Related

• Kanon vs Zod — Full comparison: philosophy, API, migration
• Zygos — Performance — Result pattern benchmarks
• Kanon Module Guide — Full module documentation