ChiKoi7/GPT-5-Distill-llama3.2-3B-Instruct-Heretic
TEXT GENERATIONConcurrency Cost:1Model Size:3.2BQuant:BF16Ctx Length:32kLicense:llama3.2Architecture:Transformer0.0K Warm

ChiKoi7/GPT-5-Distill-llama3.2-3B-Instruct-Heretic is a 3.2 billion parameter instruction-tuned language model, based on the Llama 3.2 architecture, that has been decensored using the Heretic tool. This model is a distillation of GPT-5 responses, aiming to mimic superior reasoning and tone in a lightweight package, and has been specifically processed to reduce refusals in both English and Chinese. With a 32K token context window, it is optimized for on-device chat, reasoning, summarization, and RAG applications, particularly where censorship resistance is desired.

Loading preview...