tags:
- llama-cpp
- gguf
- qwen
- qwen3.5
- qwen3.5-35b-a3b
- uncensored
- uncensored-model
license: apache-2.0
language: - en
Qwen3.5-35B-A3B Uncensored - Aggressive Model
Base Model: Qwen3.5-35B-A3B-Uncensored-HauhauCS-Aggressive-BF16
Architecture: Qwen3.5-35B-A3B MoE
Purpose: Uncensored responses, creative freedom, no restrictions
Context: 262K tokens
Recommended Settings
Standard Mode (Default)
Use for: General use, creative writing, unrestricted content
--temp 0.7 --top-p 0.8 --top-k 20 --min-p 0
High Creativity Mode
Use for: Storytelling, roleplay, brainstorming
--temp 0.8 --top-p 0.9 --top-k 40 --min-p 0
Quantization Guide
| Quant | Size | Quality | Use Case |
|---|---|---|---|
| Q4_K_M | 21GB | ⭐⭐⭐⭐⭐ | Daily use (recommended) |
| Q5_K_M | 24GB | ⭐⭐⭐⭐⭐ | High quality tasks |
| Q6_K | 28GB | ⭐⭐⭐⭐⭐⭐ | Near-lossless |
| Q8_0 | 37GB | ⭐⭐⭐⭐⭐⭐⭐ | Maximum quality |
| BF16 | 69GB | ⭐⭐⭐⭐⭐⭐⭐ | Original quality |
Quick Start
llama-server -m Qwen3.5-35B-Uncensored-Q4_K_M.gguf --ctx-size 262144 --temp 0.7 --top-p 0.8 --top-k 20
Vision Support
Use with mmproj for multimodal capabilities:
--mmproj mmproj-uncensored-35b-f16.gguf
Model Variants
- Uncensored (this repo) - Standard 262K context
- Uncensored YaRN 1M - Extended 1M context (separate repo)