Uncensored Qwen3.8-27B Derivatives Flood Hugging Face Trending Charts
From August 15-17, multiple safety-filter-removed Qwen3.8-27B derivatives trended on Hugging Face: JonathanColetti's Uncensored-GGUF, Blackfrost-AI's ABLITERATED-GGUF, and the Heretic-Abliterated-Uncensored variant reported by 0bserverx. Unsloth's quantized GGUF gained 1,000 likes within 24 hours with 1 million total downloads. The wave is driven by abliteration and low-barrier quantized deployment; an AEON-PURE variant also posted reportedly perfect benchmark scores.
Confirmed
- Three main versions: Qwen3.8-27B-Uncensored-GGUF (JonathanColetti, 08-15, llama.cpp support with speculative decoding), Qwen3.8-27B-ABLITERATED-GGUF (Blackfrost-AI, 08-16, 27B dense model, claimed image-text multimodal capability), and Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF (reported by 0bserverx, 08-17).
- Common approach: abliteration to remove safety filters and refusals, shipped in GGUF format.
- Unsloth's GGUF hit 1,000 likes in under 24 hours, ranked #3 on HF trending, 1M total downloads, runs on 17GB RAM/VRAM (per danielhanchen).
- Per StephanSturges, a community blogger tested Qwen3.8-27B-AEON-PURE scoring full marks on all "God Mode Tier" tests, claiming wins over GLM 5.2 and DeepSeek v4 Flash; it is NVFP4 natively quantized with a native version planned.
- Frizzy-MacDrizzle reported low refusal rates and good results using the Ablit model in their analytics service.
Unconfirmed
- 0bserverx noted the Heretic repo is mislabeled 3.5 but believes it is based on Qwen 3.8.
- Multimodal capability rests solely on Blackfrost-AI's description.
- The AEON-PURE perfect scores come from a single blogger with unknown test methodology.
Why it matters
- Uncensored models dominating the charts signals strong community demand for local, unrestricted generation.
- A 17GB RAM/VRAM threshold puts 27B-class models on consumer hardware; the Unsloth/GGUF/llama.cpp ecosystem is the key accelerant.
2026-08-15 ~ 2026-08-17 · 7 related posts
- Episode 1: Qwen3.8-27B Local Tests Praise It for Approaching Claude Opus(2026-08-15, 6 posts)
- Episode 2: Alibaba Open-Sources Multimodal Qwen3.8-27B as Meta Launches Muse Code(2026-08-15, 4 posts)
- Episode 3: Uncensored Qwen3.8-27B Derivatives Flood Hugging Face Trending Charts(2026-08-15, 7 posts)
- Episode 4: Qwen3.8-27B Gets INT4 and MXFP4 Quantized Releases(2026-08-16, 2 posts)
- Episode 5: Qwen3.8-27B Multimodal Model Released in GGUF Format for Local Inference(2026-08-17, 2 posts)
Primary sources
- Uncensored Qwen3.8-27B GGUF Model Trends on Hugging Face — JonathanColetti ·
- Multimodal Qwen3.8-27B-ABLITERATED-GGUF Trends on Hugging Face — Blackfrost-AI ·
- Uncensored Qwen 3.8 27B model released on Hugging Face — 0bserverx ·
- Qwen3.8-27B GGUF Hits 1,000 Likes in 24 Hours — danielhanchen · 2026-08-15
- [source] Uncensored Qwen3.8-27B GGUF Model Trends on Hugging Face — JonathanColetti · 2026-08-15
- Qwen3.8-27B-AEON-PURE scores perfect on all God Mode Tier tests — StephanSturges · 2026-08-16
- Uncensored Qwen 27B Model Released as GGUF Quantization — BLUECOW009 · 2026-08-16
- [source] Multimodal Qwen3.8-27B-ABLITERATED-GGUF Trends on Hugging Face — Blackfrost-AI · 2026-08-16
- Qwen 3.8 Ablit model available on Hugging Face — Frizzy-MacDrizzle · 2026-08-16
- [source] Uncensored Qwen 3.8 27B model released on Hugging Face — 0bserverx · 2026-08-17