Benchmarks on Public Repositories

Independent benchmarks for industry-specific AI performance.

Recently Evaluated

pallets-eco / blinker

blinker

Created10/13/2026

A fast Python in-process signal/event dispatching system.

1Model30Tasks

Languages

Python100%

Top Models

1Claude Opus 5

View Details

psf / requests

requests

Created10/13/2026

A simple, yet elegant, HTTP library.

1Model30Tasks

Languages

Python99.4%Makefile0.6%

Top Models

1Claude Opus 5

View Details

SWE-bench / SWE-bench

SWE-bench

Created10/6/2026

SWE-bench: Can Language Models Resolve Real-world Github Issues?

1Model30Tasks

Languages

Python99.3%Shell0.7%

Top Models

1Claude Opus 5

View Details

C

torvalds / linux

linux

Created8/15/2026

Linux kernel source tree

14Models85Tasks

Languages

C98%Assembly0.7%Rust0.4%Shell0.4%Python0.3%Makefile0.2%

Top Models

1GPT-5.6 Sol
2Claude Fable 5
3Claude Opus 4.8

View Details

C#

dwojtaszek / zipper

zipper

Created9/15/2026

Generate industry standard load files

2Models30Tasks

Languages

C#76.3%Shell11.6%Batchfile6.6%Python5.5%

Top Models

1GPT-5.6 Sol
2Claude Opus 5

View Details

C++

tensorflow / tensorflow

tensorflow

Created9/28/2026

An Open Source Machine Learning Framework for Everyone

5Models30Tasks

Languages

C++55.4%Python25.3%MLIR6.6%Starlark4.4%HTML4.2%Go1.2%Other2.9%

Top Models

1Claude Opus 5
2Gemini 3.6 Flash
3GPT-5.6 Luna

View Details

OpenTTD / OpenTTD

OpenTTD

Created9/22/2026

OpenTTD is an open source simulation game based upon Transport Tycoon Deluxe

2Models30Tasks

Languages

C++87.3%C9.3%CMake1.4%Squirrel1.2%Objective-C++0.6%HTML<0.1%Other0.2%

Top Models

1GPT-5.6 Sol
2Claude Opus 5

View Details

fmtlib / fmt

fmt

Created9/4/2026

A modern formatting library

2Models30Tasks

Languages

C++95.7%Python2%CMake1.7%C0.4%Shell<0.1%Cuda<0.1%Other<0.1%

Top Models

1GLM 5.2
2Muse Spark 1.2

View Details

NVIDIA / nccl

nccl

Created8/13/2026

Optimized primitives for collective multi-GPU communication

3Models30Tasks

Languages

C++53.6%Cuda18%C13.6%Python9%Cython2.3%Rust1.6%Other1.9%

Top Models

1Claude Opus 5
2GPT-5.6 Sol
3Muse Spark 1.2

View Details

Cuda

nikil-ravi / ThunderKittens

ThunderKittens

Created8/13/2026

Tile primitives for speedy kernels

1Model30Tasks

Languages

Cuda86.3%Python12.9%Makefile0.8%

Top Models

1GPT-5.6 Luna

View Details

Dart

flutter / flutter

flutter

Created8/13/2026

Flutter makes it easy and fast to build beautiful apps for mobile and beyond

3Models30Tasks

Languages

Dart75%C++16.1%Objective-C++2.7%Java2.7%Objective-C0.6%C0.6%Other2.3%

Top Models

1Gemini 3.6 Flash
2Claude Haiku 4.5 (Nonthinking)
3Gemini 3.1 Pro Preview (02/26)

View Details

Go

sashabaranov / go-openai

go-openai

Created8/21/2026

OpenAI, GPT 5.6, GPT-Image-2, Whisper API clients for Go

1Model30Tasks

Languages

Go100%

Top Models

1GPT-5.6 Sol

View Details

chrisanderton / tyk-gateway

tyk-gateway

Created8/16/2026

Tyk Open Source API Gateway written in Go, supporting REST, GraphQL, TCP and gRPC protocols

4Models30Tasks

Languages

Go98%Shell0.7%Python0.6%JavaScript0.3%Java0.2%Dockerfile<0.1%Other0.3%

Top Models

1Claude Opus 5
2GPT-5.6 Sol
3Claude Opus 4.8

View Details

ladydascalie / core

core

Created8/15/2026

A collection of packages for building a Go microservice on the LUSH platform

4Models30Tasks

Languages

Go100%

Top Models

1DeepSeek V4 Flash 0731
2GLM 5.2
3Qwen 3.7 Plus

View Details

goharbor / harbor

harbor

Created8/13/2026

An open source trusted cloud native registry project that stores, signs, and scans content.

4Models30Tasks

Languages

Go53.2%TypeScript21.4%HTML7.8%Python5.4%RobotFramework5%Jinja3.8%Other3.3%

Top Models

1Claude Opus 5
2Claude Sonnet 5
3Grok 4.5

View Details

HTML

benredmond / apex

apex

Created8/13/2026

Custom repository benchmark.

1Model30Tasks

Languages

HTML72.5%TypeScript19.3%JavaScript7.9%Shell0.2%CSS0.1%Dockerfile<0.1%

Top Models

1GPT-5.6 Luna

View Details

Java

aimacode / aima-java

aima-java

Created9/24/2026

Java implementation of algorithms from Russell And Norvig's "Artificial Intelligence - A Modern Approach"

1Model30Tasks

Languages

Java86.4%Jupyter Notebook13.6%Python<0.1%

Top Models

1Claude Haiku 4.5 (Thinking)

View Details

nationalsecurityagency / ghidra

ghidra

Created8/13/2026

Ghidra is a software reverse engineering (SRE) framework

2Models30Tasks

Languages

Java86%C++6%HTML4.1%C1.5%Python1.4%Shell0.3%Other0.6%

Top Models

1GLM 5.2
2Grok 4.5

View Details

JavaScript

vercel / next.js

next.js

Created8/15/2026

The React Framework

14Models74Tasks

Languages

JavaScript52.2%TypeScript32.1%Rust14%MDX0.8%CSS0.8%Other0.1%

Top Models

1GPT-5.6 Sol
2Kimi K3
3Claude Fable 5

View Details

Kotlin

arandomhooman / hoomans-morphe-patches

hoomans-morphe-patches

Created8/17/2026

Patches for morphe

4Models30Tasks

Languages

Kotlin89.6%Java10.4%

Top Models

1GPT-5.6 Sol
2Claude Opus 4.8
3Claude Opus 5

View Details

poalrom / sharepaste

sharepaste

Created8/14/2026

Custom repository benchmark.

2Models30Tasks

Languages

Kotlin40.5%Rust34.8%TypeScript23.6%CSS0.7%Makefile0.2%JavaScript0.1%Other<0.1%

Top Models

1Claude Opus 5
2GPT-5.6 Sol

View Details

Lean

leanprover-community / mathlib4

mathlib4

Created8/18/2026

The math library of Lean 4

4Models30Tasks

Languages

Lean99.6%Python0.3%Shell<0.1%TeX<0.1%Dockerfile<0.1%HTML<0.1%Other<0.1%

Top Models

1Claude Opus 5
2Muse Spark 1.2
3Gemini 3.6 Flash

View Details

Python

pallets-eco / blinker

blinker

Created10/13/2026

A fast Python in-process signal/event dispatching system.

1Model30Tasks

Languages

Python100%

Top Models

1Claude Opus 5

View Details

psf / requests

requests

Created10/13/2026

A simple, yet elegant, HTTP library.

1Model30Tasks

Languages

Python99.4%Makefile0.6%

Top Models

1Claude Opus 5

View Details

SWE-bench / SWE-bench

SWE-bench

Created10/6/2026

SWE-bench: Can Language Models Resolve Real-world Github Issues?

1Model30Tasks

Languages

Python99.3%Shell0.7%

Top Models

1Claude Opus 5

View Details

is42day / splink

splink

Created10/6/2026

Fast, accurate and scalable probabilistic data linkage with support for multiple SQL backends

1Model30Tasks

Languages

Python46%JavaScript40.3%Roff12.7%Jinja0.5%Shell0.2%CSS0.1%Other<0.1%

Top Models

1Claude Sonnet 5

View Details

pallets-eco / flask-wtf

flask-wtf

Created10/6/2026

Simple integration of Flask and WTForms, including CSRF, file upload and Recaptcha integration.

2Models1Task

Languages

Python100%

Top Models

1Claude Opus 5
2GPT-5.6 Luna

View Details

mkarots / benedict

benedict

Created10/1/2026

Bene(volent) dict(ator) agents for repos

1Model30Tasks

Languages

Python91.6%HTML7.5%Makefile0.9%

Top Models

1Claude Sonnet 4.6

View Details

jhonharrisricauerta2-ops / Valkyrie

Valkyrie

Created10/1/2026

Scalable, cloud-native infrastructure for evaluating AI agents across any benchmark.

1Model30Tasks

Languages

Python99.5%Makefile0.4%Dockerfile<0.1%Mako<0.1%Shell<0.1%

Top Models

1Gemini 3.6 Flash

View Details

schubydoo / clauster

clauster

Created9/30/2026

Your homelab's Claude Code launchpad. Self-hosted web UI that spawns & manages remote-control bridges on a remote host from any browser or phone — start/stop, spawn & permission modes, CLAUDE.md editor, git clone, live log tail, cost tracking. Loopback by default; password/reverse-proxy auth. No telemetry.

6Models30Tasks

Languages

Python90.2%HTML8.6%Shell0.5%PowerShell0.3%JavaScript0.2%Dockerfile<0.1%Other0.1%

Top Models

1Claude Opus 5
2GPT-5.6 Luna
3GPT-5.6 Terra

View Details

pallets / itsdangerous

itsdangerous

Created9/28/2026

Safely pass trusted data to untrusted environments and back.

1Model1Task

Languages

Python99.2%Shell0.8%

Top Models

1Claude Opus 5

View Details

vllm-project / llm-compressor

llm-compressor

Created9/23/2026

Transformers-compatible library for applying various compression algorithms to LLMs for optimized deployment with vLLM

2Models30Tasks

Languages

Python99.4%Shell0.5%Makefile<0.1%

Top Models

1Claude Opus 5
2GPT-5.6 Sol

View Details

Expertium / fsrs-autoresearch

fsrs-autoresearch

Created9/22/2026

Improving FSRS-7 with Claude

1Model30Tasks

Languages

Python92.4%Cuda6.2%C++1.3%Shell<0.1%

Top Models

1anthropic/claude-opus-5-high

View Details

JudgmentLabs / judgeval

judgeval

Created9/16/2026

The Continuous-Improvement Stack for Agents. Our environment data and evals power agent improvement and monitoring.

1Model30Tasks

Languages

Python100%

Top Models

1Claude Sonnet 5

View Details

gtrght / omnigent

omnigent

Created9/16/2026

Omnigent is an open-source AI agent framework and meta-harness: orchestrate Claude Code, Codex, Cursor, Pi, and custom agents — swap harnesses without rewriting, enforce policies and sandboxing, and collaborate in real time from any device.

1Model30Tasks

Languages

Python79.7%TypeScript17.6%JavaScript1.4%Kotlin0.3%Swift0.3%Rust0.3%Other0.5%

Top Models

1GPT-5.6 Sol

View Details

elastic / curator

curator

Created9/15/2026

Curator: Tending your Elasticsearch indices

1Model30Tasks

Languages

Python99.2%Shell0.6%Dockerfile0.2%

Top Models

1GPT-5.6 Sol

View Details

punt-labs / lux

lux

Created9/11/2026

GUI - a whiteboard and dashboard interface for agents to share visual data with humans.

3Models30Tasks

Languages

Python98.3%Shell0.8%TeX0.8%Makefile0.1%

Top Models

1anthropic/claude-opus-5-high
2Muse Spark 1.2
3GLM 5.2

View Details

avyayv / seqdataloader

seqdataloader

Created9/2/2026

Sequence data label generation and ingestion into deep learning models

1Model30Tasks

Languages

Python100%

Top Models

1GPT-5.6 Sol

View Details

django / django

django

Created9/1/2026

The Web framework for perfectionists with deadlines.

1Model30Tasks

Languages

Python97.1%Jinja1.3%JavaScript1%CSS0.5%Shell<0.1%Smarty<0.1%Other<0.1%

Top Models

1Claude Opus 5

View Details

18jeffreyma / Valkyrie

Valkyrie

Created8/18/2026

Scalable, cloud-native infrastructure for evaluating AI agents across any benchmark.

1Model30Tasks

Languages

Python99.5%Makefile0.4%Dockerfile<0.1%Mako<0.1%Shell<0.1%

Top Models

1Claude Opus 4.8

View Details

18jeffreyma / OpenHands

OpenHands

Created8/18/2026

🙌 OpenHands: Code Less, Make More

1Model30Tasks

Languages

Python77%TypeScript20.1%Shell1.4%Jinja0.9%JavaScript0.3%Makefile0.2%Other0.2%

Top Models

1Muse Spark 1.2

View Details

geopandas / geopandas

geopandas

Created8/17/2026

Python tools for geographic data

4Models30Tasks

Languages

Python100%Shell<0.1%

Top Models

1Gemini 3.6 Flash
2Claude Sonnet 5
3GPT 5.5

View Details

pallets / click

click

Created8/17/2026

Python composable command line interface toolkit

1Model30Tasks

Languages

Python100%Shell<0.1%

Top Models

1Claude Opus 5

View Details

ucbepic / docetl

docetl

Created8/17/2026

A system for agentic LLM-powered data processing and ETL

1Model30Tasks

Languages

Python68.6%TypeScript31.1%Makefile0.1%CSS<0.1%Dockerfile<0.1%JavaScript<0.1%

Top Models

1GLM 5.2

View Details

sgl-project / sglang

sglang

Created8/15/2026

SGLang is a high-performance serving framework for large language models and multimodal models.

14Models59Tasks

Languages

Python76.9%MDX6.8%Rust6.1%Cuda4.3%C++2.7%JavaScript1.9%Other1.3%

Top Models

1GPT-5.6 Terra
2Claude Fable 5
3GLM 5.2

View Details

jax-ml / jax

jax

Created8/15/2026

Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more

14Models54Tasks

Languages

Python87.1%C++9.9%Starlark2%Shell0.5%Jupyter Notebook0.4%C<0.1%

Top Models

1GPT-5.6 Sol
2Claude Fable 5
3Claude Sonnet 5

View Details

vals-ai / valkyrie

valkyrie

Created8/15/2026

Scalable, cloud-native infrastructure for evaluating AI agents across any benchmark.

14Models46Tasks

Languages

Python99.4%Makefile0.5%Mako0.1%

Top Models

1GPT-5.6 Sol
2Claude Fable 5
3Claude Opus 4.8

View Details

mirumee / ariadne

ariadne

Created8/13/2026

Python library for implementing GraphQL servers using schema-first approach. Created with :heart: by Mirumee

2Models30Tasks

Languages

Python99%HTML0.8%Just0.2%

Top Models

1GLM 5.2
2DeepSeek V4 Flash 0731

View Details

harbor-framework / terminal-bench

terminal-bench

Created8/13/2026

Measuring and evolving with the frontier of agent work

1Model30Tasks

Languages

Python48.8%JavaScript14%Lean12%TypeScript9.3%Shell4.4%Dockerfile2.2%Other9.3%

Top Models

1Claude Haiku 4.5 (Nonthinking)

View Details

skypilot-org / skypilot

skypilot

Created8/13/2026

The AI Compute Platform for frontier teams. SkyPilot turns fragmented AI compute into one AI supercomputer, so frontier AI teams build custom intelligence faster.

4Models30Tasks

Languages

Python89.4%JavaScript7.6%Jinja1.2%Shell1.1%HTML0.4%Go0.1%Other0.2%

Top Models

1Claude Opus 4.8
2Claude Opus 5
3GLM 5.2

View Details

Chen-Oliver / verifiers

verifiers

Created8/13/2026

Our library for RL environments + evals

1Model30Tasks

Languages

Python98.9%TypeScript0.7%Shell0.3%HTML<0.1%

Top Models

1Claude Opus 5

View Details

conjfrnk / memovox

memovox

Created8/13/2026

A local-first multimodal video-to-knowledge engine.

1Model30Tasks

Languages

Python97.9%JavaScript2%Makefile0.1%Dockerfile<0.1%

Top Models

1GPT-5.6 Luna

View Details

BerriAI / litellm

litellm

Created8/13/2026

The fastest, litest AI Gateway. Rust core with Python SDK. Call 100+ LLM APIs in OpenAI (or native) format with cost tracking, guardrails, load balancing, and logging [Bedrock, Azure, OpenAI, Anthropic, OpenAI, VertexAI, vLLM, Nvidia NIM]

1Model30Tasks

Languages

Python84.5%TypeScript12.6%HTML1.5%Rust0.6%HCL0.2%Go0.2%Other0.4%

Top Models

1Gemini 3.6 Flash

View Details

Ruby

rubyforgood / human-essentials

human-essentials

Created8/18/2026

Human Essentials is an inventory management system for diaper, incontinence, and period-supply banks. It supports them in distributing to partners, tracking inventory, and reporting stats and analytics.

1Model30Tasks

Languages

Ruby73.8%HTML23.6%JavaScript1.7%SCSS0.7%Shell<0.1%Dockerfile<0.1%Other<0.1%

Top Models

1DeepSeek V4 Flash 0731

View Details

Rust

AumeshMisra / tensorzero

tensorzero

Created10/3/2026

TensorZero is an open-source LLMOps platform that unifies an LLM gateway, observability, evaluation, optimization, and experimentation.

3Models30Tasks

Languages

Rust79.5%TypeScript14.7%Python4%Shell0.6%Go0.4%PLpgSQL0.3%Other0.5%

Top Models

1anthropic/claude-opus-5-high
2GPT-5.6 Luna
3GPT-5.6 Sol

View Details

Lemonawa / music163bot-rust

music163bot-rust

Created9/16/2026

Rust-powered Netease Music Bot. Built on 1% documentation and 99% pure vibes.

2Models30Tasks

Languages

Rust100%

Top Models

1GPT-5.6 Sol
2Claude Opus 5

View Details

TuringWorks / orbit-rs

orbit-rs

Created9/15/2026

Orbit-RS: A performant Multi-Protocol Data Platform

1Model30Tasks

Languages

Rust88.2%ANTLR2.9%Python2.1%PLpgSQL1.5%Shell1.3%JavaScript1.2%Other2.8%

Top Models

1Claude Haiku 4.5 (Nonthinking)

View Details

cynepiaadmin / cocoindex

cocoindex

Created8/17/2026

Data transformation framework for AI. Ultra performant, with incremental processing. 🌟 Star if you like it!

1Model30Tasks

Languages

Rust68.8%Python30.9%Shell0.2%Handlebars<0.1%

Top Models

1Claude Opus 5

View Details

astral-sh / uv

uv

Created8/15/2026

An extremely fast Python package and project manager, written in Rust.

14Models53Tasks

Languages

Rust98.2%Python1.5%Shell0.2%Other0.1%

Top Models

1Claude Fable 5
2Kimi K3
3GPT-5.6 Sol

View Details

oven-sh / bun

bun

Created8/15/2026

Incredibly fast JavaScript runtime, bundler, test runner, and package manager – all in one.

14Models48Tasks

Languages

Rust65.6%C++21.6%TypeScript8.5%C2.7%JavaScript1.3%Shell0.3%

Top Models

1Kimi K3
2Claude Fable 5
3Claude Opus 4.8

View Details

tock / tock

tock

Created8/13/2026

A secure embedded operating system for microcontrollers

2Models30Tasks

Languages

Rust97.2%Python1.4%Makefile0.7%Linker Script0.4%Shell0.2%Nix<0.1%Other<0.1%

Top Models

1Claude Sonnet 5
2DeepSeek V4 Flash 0731

View Details

Scala

apache / spark

spark

Created8/15/2026

Apache Spark - A unified analytics engine for large-scale data processing

14Models58Tasks

Languages

Scala68.5%Python16.2%Java6.7%Jupyter Notebook4.3%Other2.4%HiveQL1.9%

Top Models

1GPT-5.6 Sol
2Claude Fable 5
3Claude Opus 4.8

View Details

Shell

aeonfun / aeon

aeon

Created9/23/2026

The most autonomous AI agent framework: runs unattended on GitHub Actions, self-healing skills, drives Claude Code, Grok, Codex & more. No approval loops. Configure once, forget forever.

1Model30Tasks

Languages

Shell40.3%TypeScript36.8%JavaScript13.6%Python6.1%Solidity2.3%CSS0.9%

Top Models

1Claude Haiku 4.5 (Nonthinking)

View Details

Swift

conjfrnk / blackbird

blackbird

Created8/13/2026

A minimal, macOS-native terminal emulator.

3Models30Tasks

Languages

Swift74.2%Rust20%Shell4.5%C0.7%Metal0.2%Python0.1%Other0.3%

Top Models

1Claude Opus 5
2Muse Spark 1.2
3GPT-5.6 Luna

View Details

TypeScript

hemitoncode / DailySAT

DailySAT

Created9/24/2026

Your free SAT study app 📚 — practice tests, topic breakdowns, and tips to help every student prep with confidence.

1Model30Tasks

Languages

TypeScript91.1%JavaScript6.8%CSS1.9%Dockerfile0.2%Shell<0.1%

Top Models

1Claude Haiku 4.5 (Nonthinking)

View Details

Effect-TS / effect

effect

Created9/15/2026

Build production-ready applications in TypeScript

2Models30Tasks

Languages

TypeScript99.9%JavaScript0.1%Shell<0.1%Nix<0.1%

Top Models

1Claude Opus 5
2DeepSeek V4 Flash 0731

View Details

vastsa / PI-Desktop

PI-Desktop

Created9/10/2026

Local-first AI coding agent desktop: Electron + Rust host core + pi Agent Harness + user-installable plugins

3Models30Tasks

Languages

TypeScript57.3%Rust18.2%JavaScript17.6%CSS6%HTML0.8%Python0.1%Other<0.1%

Top Models

1Muse Spark 1.2
2DeepSeek V4 Flash 0731
3Qwen 3.7 Plus

View Details

agentspacecoding / codeburn

codeburn

Created9/1/2026

Free, local tool to track AI coding token usage and cost across 37 tools and agents (Claude Code, Cursor, Codex, Gemini and more), by model, project, and task. npx codeburn

1Model30Tasks

Languages