AnalysisDevelopersSeptember 1, 2026

slotstream runs 125B Qwen3.8-Flash-Next on 48GB Mac at ~12 tok/s

slotstream streams 4-bit Qwen3.8-Flash-Next weights from SSD via expert offloading, letting a 125B-parameter model run on Macs with as little as 16GB of RAM. Built natively with MLX and Swift, it ships an auto-mode and reports ~12 tok/s on a 48GB Mac.

1 source

More stories today

Open the live feed