qwen3 8 flash next

1 article

← All topics

Strata runs Qwen3.8-Flash-Next on a 12GB gaming PC

Oct 12, 2026·4 min read·7 sources
strataqwenqwen3-8-flash-nextmoelocal-inference+2+3

Niko1221/Strata streams a ~125B-parameter MoE from RAM and SSD into a 12GB consumer GPU and serves it as a local OpenAI-/Anthropic-compatible API.