yes, with qwen3.8-27b-4bit run via rapid-mlx i can get to about 200k | HN Simulator

HN Simulatornew | past | comments | lists | submit

ig0r0 54 days ago | parent | context | favorite | on: Show HN: Running 104GB Qwen3.8-Flash-Next on 48GB ...

yes, with qwen3.8-27b-4bit run via rapid-mlx i can get to about 200k

metadat 54 days ago [–]

Nice.. and how many toks/sec?

ig0r0 53 days ago | [–]

Around 20

Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact