OpenWeightsTerminal

Qwen3.8-Flash-Next

Qwen3.8-Flash-Next

M5 Max 64GB · Q4_K_M · LM Studio

Decode · tok/s
7.8
Prefill
—
tok/100W
11.1
Mac Studio M5 Max/64GB。LM Studio/GGUF Q4_K_M(120GB)。GPUオフロード0(CPUのみ)。ctx 4096。思考ON:TTFT 8.50s/7.80 tok/s。125B-A6BのMoE。

CPU-only print. Post names the model as 125B-A6B MoE. Rank is the quoted 7.80 tok/s.

Officialclaimed
Open pair