Top CUDA GitHub Repositories & Tools (2026)

UniLab: A Heterogeneous Architecture for Robot RL Beyond GPU-Dominant Paradigms

Persist and reuse KV Cache to speedup your LLM.

Pure Rust + CUDA LLM inference engine — no PyTorch, OpenAI-compatible, serves Qwen3 to Kimi-K2

A C++-based, cross platform ray tracing library

LargeScale Multiphysics Scientific Simulation Environment-OneFLOW CFD

Real-time multimodal desktop agent evolving toward a persistent AI OS interface (0.15 α).

Self-hosted functional language with explicit effects. Compiles to JVM bytecode, C, and NVIDIA GPU kernels via cuTile.

Local image or text → game-ready 3D on your own machine. Retopology, repaint, rigging and animation helpers, in a web viewer and a CLI. Runs Pixal3D, TRELLIS.2, Hunyuan3D and SF3D. Apple Silicon and NVIDIA. Every result carries its licence.

A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit and 4-bit floating point (FP8 and FP4) precision on Hopper, Ada and Blackwell GPUs, to provide better performance with lower memory utilization in both training and inference.

Run your AI models at home in one binary: chat, persistent memory, web access, browser control, MCP tools, encryption at rest and end-to-end encrypted remote access. Linux, macOS, Windows. FR/EN docs.

JEV-inspired parallel decisions for CUDA LLMs. One context, many decisions. vLLM API, game-agent examples, and reproducible benchmarks.

cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.

Write Java. Run on GPUs. Fast.

:metal: TT-NN operator library, and TT-Metalium low level kernel programming model.