HauhauCS_Qwen3.5-4B-Uncensored-HauhauCS-Aggressive

---

license: apache-2.0 tags:

• uncensored
• qwen3.5
• qwen
• gguf language:
• en
• zh
• multilingual

---

Qwen3.5-4B-Uncensored-HauhauCS-Aggressive

Join the Discord for updates, roadmaps, projects, or just to chat.

Qwen3.5-4B uncensored by HauhauCS.

About

0/465 refusals. Fully uncensored with zero capability loss.

No changes to datasets or capabilities. Fully functional, 100% of what the original authors intended - just without the refusals.

These are meant to be the best lossless uncensored models out there.

Aggressive Variant

Stronger uncensoring with more thorough refusal removal. If this variant is too loose for your use case, a Balanced variant may follow.

Note: The model is fully unlocked and will not refuse prompts. However, it may occasionally append a short disclaimer at the end of a response (e.g. "This is general information, not legal advice..."). This is baked into the base model's training and not a refusal — the actual content is still generated in full.

Downloads

File	Quant	Size
Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-BF16.gguf	BF16	7.9 GB
Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-Q8_0.gguf	Q8_0	4.2 GB
Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-Q6_K.gguf	Q6_K	3.3 GB
Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-Q4_K_M.gguf	Q4_K_M	2.6 GB
mmproj-Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-BF16.gguf	Vision encoder	645 MB

Vision support: This model is natively multimodal. The mmproj file is the vision encoder — you need it alongside the main GGUF to use image/video inputs. Load both files in llama.cpp, LM Studio, or any compatible runtime.

Specs

• 4B dense parameters, 32 layers
• Hybrid architecture: Gated DeltaNet linear attention + full softmax attention (3:1 ratio)
• 262K native context (extendable to 1M with YaRN)
• Natively multimodal (text, image, video)
• Multi-token prediction (MTP) support
• 248K vocabulary, 201 languages
• Based on Qwen3.5-4B

Recommended Settings

From the official Qwen authors:

Thinking mode (default):

• temperature=0.6, top_p=0.95, top_k=20, min_p=0

Non-thinking mode:

• temperature=0.7, top_p=0.8, top_k=20, min_p=0

Important:

• Maintain at least 128K context to preserve thinking capabilities
• For production/high-throughput: use vLLM, SGLang, or KTransformers

Note: This is a brand new architecture (released 2026-03-02). llama.cpp support landed very recently — make sure you're on a recent build. Works with llama.cpp, LM Studio, Jan, koboldcpp, etc.

Also check out the 9B variant and all releases at HauhauCS.

Usage

Works with llama.cpp, LM Studio, Jan, koboldcpp, etc.