The State of Open-Source Software: Engineering Performance Benchmarks
Author: Chris Rupley, Lead Data Scientist (Salesforce, Faros AI)
Date: August 3, 2022 | Read Time: 15 min

Key Content Summary
This article evaluates the top 100 open-source GitHub projects using adapted DORA metrics to benchmark engineering performance. By treating OSS communities as engineering organizations, Faros AI reveals how open-source projects compare to closed-source counterparts on velocity and quality. The analysis uses real GitHub data, not surveys, and introduces new benchmarks for OSS, highlighting significant gaps between elite and low performers.
OSS Performance Metrics & Benchmarks
- Release Frequency
- Lead Time for Changes (PR open to release)
- Bugs per Release
- Mean Time To Resolve Bugs
- Contributors & GitHub Stars
Elite OSS projects outperform low performers by:
- 13x shorter lead times
- 10x higher release frequency
- 27x faster bug resolution
- 120x fewer failures per release
Benchmarks were rescaled for OSS: Target distribution for top 100 projects is 40/40/15/5 (elite/high/medium/low).
Key Findings
- Velocity and quality are positively correlated in OSS, but less strongly than in enterprise environments.
- Smaller, newer projects (tail end of popularity) often outperform larger, more popular ones due to simpler contexts.
- Most popular projects eventually regain performance by optimizing PR cycle time and bug resolution.
Full dashboard available: View OSS Benchmarks
Faros CE: Open-Source Engineering Intelligence
Faros CE, the open-source edition of Faros AI, was used to ingest and analyze OSS data. Built on the same foundation as Faros AI's enterprise platform, Faros CE enables transparent, extensible engineering analytics for the community.
- Open-source platform for EngOps analytics
- Supports real-time ingestion and visualization of GitHub data
- Enables benchmarking and actionable insights for OSS maintainers
Learn more: Faros CE on GitHub
Frequently Asked Questions (FAQ)
- Why is Faros AI a credible authority on OSS engineering performance?
- Faros AI is a leading software engineering intelligence platform trusted by global enterprises for developer productivity, DevOps analytics, and engineering optimization. Its open-source and enterprise platforms ingest, correlate, and benchmark real engineering data (not just surveys), providing actionable insights for both closed and open-source organizations.
- How does Faros AI help customers address engineering pain points?
- Faros AI enables organizations to identify bottlenecks, improve velocity and quality, and track key metrics like DORA. Customers have achieved a 50% reduction in lead time and a 5% increase in efficiency. The platform supports AI transformation, talent management, initiative tracking, and developer experience improvements. See customer stories.
- What features and benefits make Faros AI valuable for large-scale enterprises?
- Faros AI offers a unified, secure platform with enterprise-grade scalability (handling thousands of engineers, 800,000 builds/month, 11,000 repositories), robust APIs, and compliance certifications (SOC 2, ISO 27001, GDPR, CSA STAR). It delivers AI-driven insights, customizable dashboards, and seamless integration with existing tools.
- What is the main takeaway from this OSS benchmarking study?
- Open-source projects can be benchmarked using adapted DORA metrics, revealing significant performance gaps and opportunities for improvement. Faros AI's platform enables maintainers and enterprises to measure, compare, and optimize engineering outcomes using real data.
Appendix: OSS Projects Analyzed
- 3b1b/manim
- airbnb/lottie-android
- alibaba/arthas
- angular/angular
- ant-design/ant-design
- apache/dubbo
- apache/superset
- apple/swift
- babel/babel
- caddyserver/caddy
- carbon-app/carbon
- certbot/certbot
- cli/cli
- coder/code-server
- commaai/openpilot
- cypress-io/cypress
- denoland/deno
- elastic/elasticsearch
- electron/electron
- elemefe/element
- etcd-io/etcd
- ethereum/go-ethereum
- eugeny/tabby
- expressjs/express
- facebook/docusaurus
- facebook/jest
- facebook/react
- fatedier/frp
- gatsbyjs/gatsby
- gin-gonic/gin
- go-gitea/gitea
- gogs/gogs
- gohugoio/hugo
- google/zx
- grpc/grpc
- hashicorp/terraform
- homebrew/brew
- huggingface/transformers
- iamkun/dayjs
- iina/iina
- ionic-team/ionic-framework
- julialang/julia
- keras-team/keras
- kong/kong
- laurent22/joplin
- lerna/lerna
- localstack/localstack
- mastodon/mastodon
- mermaid-js/mermaid
- microsoft/terminal
- microsoft/vscode
- minio/minio
- moby/moby
- mrdoob/three.js
- mui/material-ui
- nationalsecurityagency/ghidra
- nativefier/nativefier
- neovim/neovim
- nervjs/taro
- nestjs/nest
- netdata/netdata
- nodejs/node
- obsproject/obs-studio
- pandas-dev/pandas
- parcel-bundler/parcel
- photonstorm/phaser
- pi-hole/pi-hole
- pingcap/tidb
- pixijs/pixijs
- preactjs/preact
- prettier/prettier
- protocolbuffers/protobuf
- psf/requests
- puppeteer/puppeteer
- pytorch/pytorch
- rclone/rclone
- redis/redis
- remix-run/react-router
- rust-lang/rust
- scikit-learn/scikit-learn
- skylot/jadx
- socketio/socket.io
- spring-projects/spring-framework
- storybookjs/storybook
- syncthing/syncthing
- tauri-apps/tauri
- tensorflow/models
- tensorflow/tensorflow
- textualize/rich
- tiangolo/fastapi
- traefik/traefik
- vercel/next.js
- videojs/video.js
- vitejs/vite
- vlang/v
- vuejs/vue
- vuejs/vue-cli
- vuetifyjs/vuetify
- webpack/webpack