What are the hardware requirements for Qwen3.8 27B FULL NVFP4?

Qwen3.8 27B FULL NVFP4 requires approximately 18.8 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Qwen3.8 27B FULL NVFP4 and how large is it?

Qwen3.8 27B FULL NVFP4 is developed by digi-texx, a model with 27b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Qwen3.8 27B FULL NVFP4?

Qwen3.8 27B FULL NVFP4 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Qwen3.8 27B FULL NVFP4 by digi-texx — VRAM 18.8GB

Name: Qwen3.8 27B FULL NVFP4
Author: digi-texx

Qwen3.8 27B FULL NVFP4 is an open-source language model by digi-texx. Features: 27b LLM, VRAM: 18.8GB, Instruction-Based, LLM Explorer Score: 0.32.

Arxiv:2509.23202 Arxiv:2608.13966 Arxiv:2608.28113 8-bit Agentic Base model:quantized:quasar-qa... Base model:quasar-qat/qwen3.8-... Compressed-tensors Conversational Dataset:digi-texx/calib-agenti... Dataset:digi-texx/calib-agenti... Dataset:digi-texx/calib-agenti... Dataset:glaiveai/glaive-functi... Dataset:nebius/swe-rebench-ope... Dataset:nvidia/nemotron-math-v... Dataset:nvidia/nemotron-sft-ag... Dataset:nvidia/nemotron-sft-in... Dataset:nvidia/nemotron-sft-op... Dataset:nvidia/nemotron-sft-sw... Dataset:salesforce/apigen-mt-5... Dataset:salesforce/xlam-functi... Dataset:swe-bench/swe-smith-tr... Dataset:team-ace/toolace En Endpoints compatible Fp4 H-scale Image-text-to-text Instruct Mr-gptq Nvfp4 Post-training-quantization Quantization Qwen3 5 Region:us Safetensors Sharded Tensorflow Tool-calling W4a4

Model Card on HF 🤗: https://huggingface.co/digi-texx/Qwen3.8-27B-FULL-NVFP4

Qwen3.8 27B FULL NVFP4 Parameters and Internals

LLM Name	Qwen3.8 27B FULL NVFP4
Repository 🤗	https://huggingface.co/digi-texx/Qwen3.8-27B-FULL-NVFP4
Base Model(s)	QUASAR-QAT/Qwen3.8-27B-QUASAR-NVFP4 QUASAR-QAT/Qwen3.8-27B-QUASAR-NVFP4
Model Size	27b
Required VRAM	18.8 GB
Updated	2026-09-05
Maintainer	digi-texx
Model Type	qwen3_5
Instruction-Based	Yes
Model Files	5.0 GB: 1-of-5 5.0 GB: 2-of-5 5.0 GB: 3-of-5 3.1 GB: 4-of-5 0.7 GB: 5-of-5
Supported Languages	en
Model Architecture	Qwen3_5ForConditionalGeneration
Model Max Length	262144
Transformers Version	5.8.0.dev0
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Torch Data Type	bfloat16
Errors	replace

Rank the Qwen3.8 27B FULL NVFP4 capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Qwen3.8 27B FULL NVFP4

Best Alternatives	Context / RAM	Downloads	Likes
Qwopus3.8 27B Flash	0K / 55.6 GB	234	19
Tmax 27B	0K / 53.8 GB	4630	26
...8 27B Brainwaves WFH Mxfp4 Mlx	0K / 15.1 GB	490	1
...sion GAIN V1.1 Fable Mxfp8 Mlx	0K / 28.7 GB	720	2
...1 Darker Hero GAIN B Mxfp8 Mlx	0K / 28.7 GB	609	2
... 27B Brainwaves 2M Qx64 Hi Mlx	0K / 21 GB	552	1
... 27B Brainwaves 1M Qx86 Hi Mlx	0K / 27 GB	427	1
Qwen3.8 27B Brainwaves WFH	0K / 55.5 GB	174	1
Qwen3.8 27B Brainwaves	0K / 55.5 GB	401	1
...en3.8 27B Brainwaves Mxfp8 Mlx	0K / 28.7 GB	361	1

Note: green Score (e.g. "73.2") means that the model is better than digi-texx/Qwen3.8-27B-FULL-NVFP4.