Intelligent Mixture-of-Models Router for Efficient LLM Inference

A community-maintained Python framework for creating mathematical animations.

A Datacenter Scale Distributed Inference Serving Framework

FlashInfer: Kernel Library for LLM Serving

A Distributed Attention Towards Linear Scalability for Ultra-Long Context, Heterogeneous Data Training

Transform data, train models, and run SQL with marimo — feels like a next-gen reactive notebook, stored as Git-friendly Python. Deploy as scripts, pipelines, endpoints, and apps. All from an AI-nat…

CUDA Templates for Linear Algebra Subroutines

An implementation of the 5th Edition game system for Foundry Virtual Tabletop (http://foundryvtt.com).

TensorRT-LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and build TensorRT engines that contain state-of-the-art optimizations to perform inference efficie…

PyTorch code and models for V-JEPA self-supervised learning from video.

Implementation for FP8/INT8 Rollout for RL training without performence drop.

slime is a LLM post-training framework aiming for RL Scaling.

Frontier Models playing the board game Diplomacy.

veRL: Volcano Engine Reinforcement Learning for LLM

SGLang is a structured generation language designed for large language models (LLMs). It makes your interaction with models faster and more controllable.

An open-source AI agent that brings the power of Gemini directly into your terminal.

AI coding agent, built for the terminal.

Qwen3-Coder is the code version of Qwen3, the large language model series developed by Qwen team, Alibaba Cloud.

Train speculative decoding models effortlessly and port them smoothly to SGLang serving.

Fully open reproduction of DeepSeek-R1

MAGI-1: Autoregressive Video Generation at Scale

Sampling profiler for Python programs

Humanity's Last Exam

Development repository for the Triton language and compiler

Lightweight coding agent that runs in your terminal

An AI Hedge Fund Team

AutoAWQ implements the AWQ algorithm for 4-bit quantization with a 2x speedup during inference. Documentation:

GitHub's official MCP Server

About Awesome things towards foundation agents. Papers / Repos / Blogs / ...