Qwen3-4B Heretic Z-Image V2 (Merged)

Single-file merged version of Lockout/qwen3-4b-heretic-zimage V2.

What is this?

This is the actual Z-Image trained Qwen encoder with abliteration applied using the Heretic tool.

Key points:

Uses the encoder that was trained alongside Z-Image-Turbo (not a random Qwen)
V2 has lower KLD (less model degradation from abliteration)
Prevents "refusal embeddings" that cause unwanted clothing/covering in generations

Drop-in replacement for qwen_3_4b.safetensors in ComfyUI:

ComfyUI/models/text_encoders/qwen3_4b_heretic_zimage_v2.safetensors

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Base model

Finetuned

(1)

this model