microsoft/BitNet
One-bit LLM inference framework
An official inference framework and runtime for running one-bit large language models efficiently on CPUs. It provides optimized C++ and Python implementations for fast local inference of quantized models.
- Language
- C++
- Stars
- 40,800
- Contributors
- 23
- Forks
- 3,800
- Days trending in the last 30
- 1 day
- Commits
- 110
- Last commit
- July 27, 2026