hugpadhugpad

unsloth / Qwen3.8-27B-NVFP4

Other20B1.6M530Updated Sep 8, 2026
Hugging Face

Downloads

1.6M

Likes

530

Last updated

Sep 8, 2026

License

apache-2.0

[!Note] This model works with vLLM. It also works with SGLang, but only with the v0.5.19, because the lmhead is FP8 rather than 4-bit.

Tags

  • qwen3_5
  • unsloth
  • compressed-tensors

Technical specifications

Parameters20B
LibraryNot available
LanguagesNot available
Base modelQwen/Qwen3.8-27B
Licenseapache-2.0