LLM EXPLORER 63,197 MODELS INDEXED

Qwen3.8 Flash Next P48NVFP4 MoESQ by ISTA-DASLab

By ISTA-DASLab · 738 downloads
HF Repo ↗ Copy URL

Qwen3.8 Flash Next P48NVFP4 MoESQ is an open-source language model by ISTA-DASLab. Features: 134.7b LLM, VRAM: 152.7GB, License: other, LLM Explorer Score: 0.35.

  Arxiv:2610.02241   8-bit   Activation-quantization Base model:quantized:qwen/qwen... Base model:qwen/qwen3.8-flash-...   Blackwell   Compressed-tensors   Conversational   En   Endpoints compatible   Image-text-to-text   Moe   Nvfp4   Paired-4:8   Qwen4-exp   Qwen4 exp   Region:us   Safetensors   Semi-structured-sparsity   Sharded   Sparse-storage   Tensorflow   W4a4
Model Card on HF 🤗: https://huggingface.co/ISTA-DASLab/Qwen3.8-Flash-Next-P48NVFP4-MoESQ 

Qwen3.8 Flash Next P48NVFP4 MoESQ Parameters and Internals

LLM NameQwen3.8 Flash Next P48NVFP4 MoESQ
Repository 🤗https://huggingface.co/ISTA-DASLab/Qwen3.8-Flash-Next-P48NVFP4-MoESQ 
Base Model(s)  Qwen/Qwen3.8-Flash-Next   Qwen/Qwen3.8-Flash-Next
Model Size134.7b
Required VRAM152.7 GB
Updated2026-10-06
MaintainerISTA-DASLab
Model Typeqwen4_exp
Model Files  3.5 GB: 1-of-49   6.2 GB: 2-of-49   103.0 GB: 3-of-49   1.0 GB: 4-of-49   1.0 GB: 5-of-49   1.0 GB: 6-of-49   1.0 GB: 7-of-49   1.0 GB: 8-of-49   1.0 GB: 9-of-49   1.0 GB: 10-of-49   1.0 GB: 11-of-49   1.0 GB: 12-of-49   1.0 GB: 13-of-49   1.0 GB: 14-of-49   1.0 GB: 15-of-49   1.0 GB: 16-of-49   1.0 GB: 17-of-49   1.0 GB: 18-of-49   1.0 GB: 19-of-49   1.0 GB: 20-of-49   1.0 GB: 21-of-49   1.0 GB: 22-of-49   1.0 GB: 23-of-49   1.0 GB: 24-of-49   1.0 GB: 25-of-49   1.0 GB: 26-of-49   1.0 GB: 27-of-49   1.0 GB: 28-of-49   1.0 GB: 29-of-49   1.0 GB: 30-of-49   1.0 GB: 31-of-49   1.0 GB: 32-of-49   1.0 GB: 33-of-49   1.0 GB: 34-of-49   1.0 GB: 35-of-49   1.0 GB: 36-of-49   1.0 GB: 37-of-49   1.0 GB: 38-of-49   1.0 GB: 39-of-49   1.0 GB: 40-of-49   1.0 GB: 41-of-49   1.0 GB: 42-of-49   1.0 GB: 43-of-49
Supported Languagesen
Model ArchitectureQwen4ExpForConditionalGeneration
Licenseother
Transformers Version5.8.0.dev0