Issue 2026-10-30 · Industry · 开源 · 模型发布 · 研究
Community releases Qwen 3.8 Swift 1.5 Flash Heretic v2 uncensored quant
Community developers released an uncensored Heretic v2 GGUF quant of Qwen 3.8 Swift 1.5 Flash, tuned for reverse engineering and coding, with bit allocation weighted toward chain-of-thought, system prompts and architecture. The authors say reasoning and output length are much improved over v1.
Read original ↗- Developer builds Halloween haunted house with NVIDIA Nemotron models
- Community releases Gemma-4-31B-Humanlike-Chat fine-tune
- Whissle Open-Sources 70MB Arc Streaming Speech Model
- Hugging Face and partners launch ProtoStar to make public science AI-ready
- Hugging Face launches AnySpeechRecognizer and AnyDecisionModel for Apple platforms