Laya Mlx

Local LLM Testing & Benchmarking for Apple Silicon

Gemma 4 26B-A4B inference in ~2 GB of RAM on any M-series MacBook

The fastest way to run Qwen 3.8 Flash Next, Qwen 3.8 27B and Ternary Bonsai 2 27B on a Mac: 125 tok/s in OpenCode on an M5 Max, and a 27B model on 16 GB Macs.

The open-source System One decision model. Sub-15ms, non-autoregressive, local drop-in alternative to TypeSafe Jev.

Run a 105 GB AI model on a Mac that can't hold it. Slotstream streams Qwen3.8-Flash-Next (125B mixture of experts) from your SSD and caches the busiest experts…

An easy-to-use, fast toolkit to scale up RL post-training on a single node.