What are the hardware requirements for Qwen3.8 27B USS Prism?

Qwen3.8 27B USS Prism requires approximately 55.5 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Qwen3.8 27B USS Prism and how large is it?

Qwen3.8 27B USS Prism is developed by nightmedia, 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 USS Prism?

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

Qwen3.8 27B USS Prism by nightmedia — VRAM 55.5GB

Name: Qwen3.8 27B USS Prism
Author: nightmedia

Qwen3.8 27B USS Prism is an open-source language model by nightmedia. Features: 27b LLM, VRAM: 55.5GB, License: apache-2.0, Instruction-Based, Code Generating, Merged.

Merged Model 1m context 256k context All genres All use cases Base model:adapter:jackrong/qw... Base model:jackrong/qwopus3.8-... Bf16 Chain-of-thought Claude-distillation Claude4.6 Codegen Coding Conversational Creative Creative writing Deckard(qx) Distillation En Endpoints compatible Es Experimental Fable Fiction Fiction story Fiction writing Image-text-to-text Instruct Instruction-tuned Ja Long-cot Lora Math Mlx Multilingual Plot generation Polaris Polaris-alpha Qwen Qwen3.5 Qwen3.6 Qwen3 5 Qwen3 6 Reasoning Region:us Research Roleplaying Safetensors Scene continue Science fiction Sft Sharded Stem Story Story generation Storytelling Sub-plot generation Tensorflow Unsloth Vivid prosing Vivid writing Writing Zh

Model Card on HF 🤗: https://huggingface.co/nightmedia/Qwen3.8-27B-USS-Prism

Qwen3.8 27B USS Prism Parameters and Internals

LLM Name	Qwen3.8 27B USS Prism
Repository 🤗	https://huggingface.co/nightmedia/Qwen3.8-27B-USS-Prism
Base Model(s)	Jackrong/Qwopus3.8-27B-Coder-Flash B1 27B nightmedia/Qwen3.6-27B-Seven nightmedia/Qwen3.8-27B-Brainwaves nightmedia/Qwen3.8-27B-MindMeld OmniJev/OneJev1.6-27B Jackrong/Qwopus3.8-27B-Coder-Flash schneewolflabs/B1-27B nightmedia/Qwen3.6-27B-Seven nightmedia/Qwen3.8-27B-Brainwaves nightmedia/Qwen3.8-27B-MindMeld OmniJev/OneJev1.6-27B
Merged Model	Yes
Model Size	27b
Required VRAM	55.5 GB
Updated	2026-11-04
Maintainer	nightmedia
Model Type	qwen3_5
Instruction-Based	Yes
Model Files	2.5 GB: 1-of-12 4.8 GB: 2-of-12 5.0 GB: 3-of-12 4.9 GB: 4-of-12 5.0 GB: 5-of-12 4.9 GB: 6-of-12 5.0 GB: 7-of-12 4.9 GB: 8-of-12 5.0 GB: 9-of-12 4.9 GB: 10-of-12 4.9 GB: 11-of-12 3.7 GB: 12-of-12
Supported Languages	en zh ja es
Generates Code	Yes
Model Architecture	Qwen3_5ForConditionalGeneration
License	apache-2.0
Model Max Length	262144
Tokenizer Class	TokenizersBackend
Padding Token	<\|endoftext\|>
Errors	replace

Rank the Qwen3.8 27B USS Prism 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