Abliterated Qwen3.8-27B Variants Released on Hugging Face

Huihui-ai released an abliterated fine-tune of Qwen3.8-27B on Hugging Face, followed by an MLX-based multimodal variant with image-to-text support and low-bit quantization options for local deployment.

2026-08-17 ~ 2026-08-18 · 2 related posts

• Qwen3.8-27B-Abliterated-MLX Model Released on Hugging Face — PocketAiHub · 2026-08-17
• Huihui releases uncensored Qwen3.8-27B abliterated model — huihui-ai · 2026-08-18