Open-Source Laya Ported to Apple MLX for 7ms On-Device Decisions

Developer mizorewww ported the open-source decision model Laya to Apple MLX, enabling fully local gameplay on a 16GB MacBook Air with under 1GB peak memory and decisions as fast as 7.39ms—11x faster than cloud-based Jev, with zero token usage.

2026-09-20 ~ 2026-09-22 · 3 related posts

• Laya-MLX: open typed decision model runs locally at 7–14 ms per decision, zero tokens — JiliJeanlouis · 2026-09-20
• Open-weights Laya plays Tetris locally, deciding 11x faster than cloud-based Jev on a MacBook Air — alexcovo_eth · 2026-09-22
• Laya-MLX Runs Open-Source Decision Model Locally on Mac in Under 1GB, 7.39ms per Decision — 量子位 · 2026-09-22