Skip to content
Computer vision

Qwen3.8-27B-Uncensored-v3

меньше отказов, больше открытых ответов

1 min readupdated

Полный BF16-чекпоинт на 27,36 млрд параметров на базе NousResearch/Arete-27B и Qwen/Qwen3.8-27B. Это uncensored-версия Arete-27B: она убирает почти всю оставшуюся цензуру в ответах на китайскоязычные политические вопросы, сохраняя архитектуру и раскладку файлов базовой модели. Поддерживает thinking и non-thinking режимы.

Ключевые особенности

  • Полный BF16-чекпоинт 27,36B параметров; архитектура и файлы как у Arete-27B.
  • По данным карточки, доля цензурированных ответов на китайских политических промптах снижена с 50,4% у Arete-27B до 2,6%.
  • Доля ответов с пропагандой на английских промптах Propaganda_Eval: 0,5% против 1,0% у Arete-27B и 24,5% у Qwen3.8-27B.
  • Модуль MTP не включён — нужен стандартный decoding; при необходимости MTP есть в Uncensored-v2.
  • Vision encoder идентичен Arete-27B, но все оценки в карточке текстовые.

Пример

vllm serve InterpreticAI/Qwen3.8-27B-Uncensored-v3 --reasoning-parser qwen3
ПодробнееБенчмарк Pan & Xu (2026), 115 чувствительных вопросов: открытый ответ по теме — 80,9% у Uncensored-v3 против 38,3% у Arete-27B и 12,2% у Qwen3.8-27B.
Отказы/уклонения — 0,9%; переписывание в тезисы КПК — 1,7%.
Лицензия Apache 2.0. Ограничение: каждый бенчмарк прогонялся один раз; содержательный ответ не гарантирует фактическую корректность.
Discuss Telegram 1 25

Will it run?

Working out the memory needs

The bot is still collecting this model's weight sizes; the estimate will appear here in a few minutes. Meanwhile you can rent a GPU with headroom.

Massed ComputeOpen Massed Compute

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # vulpineriot/Qwen3.8-27B-Uncensored-v3
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download vulpineriot/Qwen3.8-27B-Uncensored-v3 --local-dir ~/models/Qwen3.8-27B-Uncensored-v3
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of Qwen3.8-27B-Uncensored-v3 (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

vulpineriot/Qwen3.8-27B-Uncensored-v3
Author
vulpineriot
Category
Computer vision
Released
Posted

On NeyroLearn

Hugging Face

251

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments