LLM EXPLORER 62,939 MODELS INDEXED

Qwen3.5 397B A17B MoE MXFP4 by amd

By amd · 19757 downloads
HF Repo ↗ Copy URL

Qwen3.5 397B A17B MoE MXFP4 is an open-source language model by amd. Features: 221.8b LLM, VRAM: 98.9GB, License: apache-2.0, MoE, LLM Explorer Score: 0.31.

  8-bit Base model:quantized:qwen/qwen... Base model:qwen/qwen3.5-397b-a...   Conversational   En   Endpoints compatible   Image-text-to-text   Moe   Quark   Qwen3 5 moe   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/amd/Qwen3.5-397B-A17B-MoE-MXFP4 

Qwen3.5 397B A17B MoE MXFP4 Parameters and Internals

LLM NameQwen3.5 397B A17B MoE MXFP4
Repository 🤗https://huggingface.co/amd/Qwen3.5-397B-A17B-MoE-MXFP4 
Base Model(s)  Qwen/Qwen3.5-397B-A17B-FP8   Qwen/Qwen3.5-397B-A17B-FP8
Model Size221.8b
Required VRAM98.9 GB
Updated2026-08-09
Maintaineramd
Model Typeqwen3_5_moe
Model Files  2.3 GB: 1-of-94   2.3 GB: 2-of-94   2.3 GB: 3-of-94   2.3 GB: 4-of-94   2.3 GB: 5-of-94   2.3 GB: 6-of-94   2.3 GB: 7-of-94   2.3 GB: 8-of-94   2.3 GB: 9-of-94   2.3 GB: 10-of-94   2.3 GB: 11-of-94   2.3 GB: 12-of-94   2.3 GB: 13-of-94   2.3 GB: 14-of-94   2.3 GB: 15-of-94   2.3 GB: 16-of-94   2.3 GB: 17-of-94   2.3 GB: 18-of-94   2.3 GB: 19-of-94   2.3 GB: 20-of-94   2.3 GB: 21-of-94   2.3 GB: 22-of-94   2.3 GB: 23-of-94   2.3 GB: 24-of-94   2.3 GB: 25-of-94   2.3 GB: 26-of-94   2.3 GB: 27-of-94   2.3 GB: 28-of-94   2.3 GB: 29-of-94   2.3 GB: 30-of-94   2.3 GB: 31-of-94   2.3 GB: 32-of-94   2.3 GB: 33-of-94   2.3 GB: 34-of-94   2.3 GB: 35-of-94   2.3 GB: 36-of-94   2.3 GB: 37-of-94   2.3 GB: 38-of-94   2.3 GB: 39-of-94   2.3 GB: 40-of-94   2.3 GB: 41-of-94   2.3 GB: 42-of-94   2.3 GB: 43-of-94
Supported Languagesen
Model ArchitectureQwen3_5MoeForConditionalGeneration
Licenseapache-2.0
Transformers Version4.57.0.dev0