Skip to main content
These benchmarks compare the performance of Fiber against other web frameworks using industry-standard tests. The results demonstrate Fiber’s exceptional performance characteristics across various scenarios.

​
TechEmpower Benchmarks

TechEmpower provides a comprehensive performance comparison of web application frameworks executing fundamental tasks such as JSON serialization, database access, and server-side template rendering. Each framework operates under a realistic production configuration, with results recorded on both cloud instances and physical hardware. Test implementations are community-contributed and maintained in the FrameworkBenchmarks repository.

​
Test Environment

​
Hardware Specifications

  • CPU: 56 Cores Intel(R) Xeon(R) Gold 6330 @ 2.00GHz
  • Hardware: Three homogeneous ProLiant DL360 Gen10 Plus
  • RAM: 64GB
  • Storage: Enterprise SSD
  • OS: Ubuntu
  • Network: Mellanox Technologies MT28908 Family ConnectX-6 40Gbps Ethernet
  • Fiber Version: v3.0.0

​
Benchmark Results

​
Plaintext

The Plaintext test measures basic request routing and demonstrates the capacity of high-performance platforms. Requests are pipelined, and the tiny response body demands high throughput to saturate the benchmark’s gigabit Ethernet. See Plaintext requirements

Fiber

11,987,976 responses per secondAverage latency: 1.0 ms

Express

1,204,969 responses per secondAverage latency: 8.8 ms
Fiber handles 10x more requests than Express in plaintext benchmarks while maintaining significantly lower latency.
Plaintext Performance Fiber vs Express Plaintext

​
JSON Serialization

The JSON test measures JSON serialization performance, a common operation in modern web APIs.

Fiber

2,363,294 responses per secondAverage latency: 0.2 ms

Express

949,717 responses per secondAverage latency: 0.5 ms
JSON Performance Fiber vs Express JSON

​
Single Query

The Single Query test exercises the framework’s object-relational mapping (ORM) or database access layer with a single database query per request.

Fiber

953,016 responses per secondAverage latency: 0.6 ms

Express

441,543 responses per secondAverage latency: 1.3 ms
Single Query Performance Fiber vs Express Single Query

​
Multiple Queries

The Multiple Queries test performs multiple database queries per request, testing how frameworks handle concurrent database access.

Fiber

54,002 responses per secondAverage latency: 9.4 ms

Express

85,011 responses per secondAverage latency: 6.0 ms
Multiple Queries Performance Fiber vs Express Multiple Queries

​
Data Updates

The Data Updates test performs database queries and updates, representing a more complete CRUD operation pattern.

Fiber

29,984 responses per secondAverage latency: 16.9 ms

Express

54,887 responses per secondAverage latency: 9.2 ms
Data Updates Performance Fiber vs Express Data Updates

​
Performance Analysis

​
Where Fiber Excels

Fiber demonstrates exceptional performance in high-throughput scenarios:
  • Plaintext: 10x faster than Express (11.9M vs 1.2M req/s)
  • JSON Serialization: 2.5x faster than Express (2.4M vs 950K req/s)
  • Single Query: 2.2x faster than Express (953K vs 442K req/s)
This makes Fiber ideal for:
  • High-traffic APIs
  • Microservices handling many requests
  • Real-time applications
  • Gateway and proxy services
Fiber maintains consistently low latency across most operations:
  • Plaintext: 1.0ms average latency
  • JSON: 0.2ms average latency
  • Single Query: 0.6ms average latency
Low latency is crucial for:
  • User-facing applications
  • API gateways
  • Service mesh communications
Fiber’s zero-allocation router and efficient memory management contribute to:
  • Lower memory footprint
  • Better garbage collection performance
  • Higher request throughput under memory pressure
  • Reduced infrastructure costs

​
Benchmark Methodology

​
Test Categories

TechEmpower benchmarks test several key aspects of web framework performance:
  1. JSON Serialization: Tests the framework’s ability to serialize objects to JSON
  2. Single Database Query: Tests database access with a single query per request
  3. Multiple Database Queries: Tests handling of multiple sequential database queries
  4. Database Updates: Tests reading, modifying, and updating database records
  5. Plaintext: Tests basic request routing and response handling
  6. Fortunes: Tests template rendering and HTML escaping

​
Why These Tests Matter

Real-World Patterns

Tests reflect common web application patterns: JSON APIs, database access, and content rendering.

Standardized Comparison

All frameworks run identical operations under the same hardware conditions for fair comparison.

Production Configuration

Frameworks are configured as they would be in production environments.

Community Validated

Test implementations are reviewed and maintained by the community.

​
Interpreting Results

Benchmarks measure specific scenarios and may not represent your application’s exact workload. Use these results as a general guide, and always benchmark your specific use case.

​
Considerations

  • Database tests: Performance depends heavily on database configuration and network latency
  • Real applications: Include business logic, validation, and other processing not captured in benchmarks
  • Scalability: Both vertical (single machine) and horizontal (multiple machines) scaling characteristics matter
  • Development velocity: Framework productivity, ecosystem, and maintainability are crucial factors beyond raw performance

​
Running Your Own Benchmarks

To benchmark Fiber in your environment: 
# Clone the TechEmpower benchmarks repository
git clone https://github.com/TechEmpower/FrameworkBenchmarks.git
cd FrameworkBenchmarks

# Run Fiber benchmarks
./tfb --mode benchmark --test fiber
For custom benchmarks specific to your use case: 
package main

import (
    "testing"
    "github.com/gofiber/fiber/v3"
)

func BenchmarkFiberHandler(b *testing.B) {
    app := fiber.New()
    app.Get("/", func(c fiber.Ctx) error {
        return c.SendString("Hello, World!")
    })
    
    b.ResetTimer()
    b.RunParallel(func(pb *testing.PB) {
        for pb.Next() {
            // Your benchmark code
        }
    })
}
For production benchmarking, use realistic data volumes, network conditions, and concurrent user patterns that match your expected traffic.

Build docs developers (and LLMs) love

Get started for freeTalk to us