Uncensored Qwen3.6-27B Derivative 'Fable-Fusion-711' Lands on Hugging Face With Multimodal Support and Record ARC-C Score

A creator known as DavidAU has released an abliterated, uncensored fine-tune of Alibaba's Qwen3.6-27B on Hugging Face, stripping out safety refusals while retaining full image-text-to-text multimodal capability and posting benchmark scores that edge past the original base model. The model, awkwardly named Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF, was published around July 18-19, 2026, and packaged as GGUF quantizations intended to run locally on consumer hardware.