Skip to content

Releases

Molten v0.1.0

Molten v0.1.0 Latest
Latest

Choose a tag to compare

View all tags
@zhangshuibai zhangshuibai released this 25 Oct 15:21
v0.1.0
2c54857

First release of Molten, a native Rust + Metal inference engine for LiquidAI LFM2.5-24B-A2B (MLX 4-bit) on Apple M5-series Macs: 300+ tok/s decode on an M5 Max, 1.45-1.6x mlx-lm and 1.3-1.5x llama.cpp from 256 to 128K tokens of context.

Install

curl -L https://github.com/zhangshuibai/molten/releases/latest/download/molten-macos-arm64.tar.gz | tar xz
cd molten-macos-arm64
huggingface-cli download LiquidAI/LFM2.5-24B-A2B-MLX-4bit \
  --revision 667408d494bde191c2cb9dd4f4549827816b9275 --local-dir ~/models/lfm2.5
./molten chat --model ~/models/lfm2.5

Requirements: an M5, M5 Pro or M5 Max Mac with 24 GB or more of memory and macOS 26 or newer. If you download the archive with a web browser, run xattr -d com.apple.quarantine molten first.

Assets

  • molten-macos-arm64.tar.gz: the molten executable (arm64, ad-hoc signed, not notarized) with LICENSE and NOTICE. Built with cargo build --release --locked (Rust 1.92.0) from the source of the v0.1.0 tag.
  • molten-macos-arm64.tar.gz.sha256: SHA-256 db255d9f1c293eb856d5788c1f80b3ae00261569f0962ee780b70cab9d9611b0.
Assets 4
Loading