LLM EXPLORER 58,608 MODELS INDEXED

Qwen3.8 27B Quark Qronos INT4 W4A16 by amd

By amd · 413 downloads
HF Repo ↗ Copy URL

Qwen3.8 27B Quark Qronos INT4 W4A16 is an open-source language model by amd. Features: 27b LLM, VRAM: 19.9GB, License: apache-2.0.

  Arxiv:2505.11695   Amd Base model:quantized:qwen/qwen...   Base model:qwen/qwen3.8-27b   Conversational   Endpoints compatible   Image-text-to-text   Int4   Qronos   Quantization   Quark   Qwen3 5   Region:us   Safetensors   W4a16
Model Card on HF 🤗: https://huggingface.co/amd/Qwen3.8-27B-Quark-Qronos-INT4-W4A16 

Qwen3.8 27B Quark Qronos INT4 W4A16 Parameters and Internals

LLM NameQwen3.8 27B Quark Qronos INT4 W4A16
Repository 🤗https://huggingface.co/amd/Qwen3.8-27B-Quark-Qronos-INT4-W4A16 
Base Model(s)  Qwen/Qwen3.8-27B   Qwen/Qwen3.8-27B
Model Size27b
Required VRAM19.9 GB
Updated2026-08-17
Maintaineramd
Model Typeqwen3_5
Model Files  19.9 GB
Model ArchitectureQwen3_5ForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Transformers Version5.2.0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Errorsreplace

Best Alternatives to Qwen3.8 27B Quark Qronos INT4 W4A16

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3.8 27B FP80K / 0.4 GB123157416
Qwen3.8 27B0K / 55.6 GB28753
Qwen3.6 27B0K / 55.4 GB67281382197
Qwen3.6 27B NVFP40K / 22 GB1887525419
...n3.8 27B MTPLX Optimized Speed0K / 19.6 GB698950
ThinkingCap Qwen3.6 27B0K / 54.7 GB35947614
Qwen3.6 27B FP80K / 0.4 GB7740472342
Qwen3.5 27B0K / 55.2 GB25362941028
Qwen3.5 27B0K / 55.2 GB811516
Qwen3.8 27B NVFP40K / 22 GB3012214
Note: green Score (e.g. "73.2") means that the model is better than amd/Qwen3.8-27B-Quark-Qronos-INT4-W4A16.