Professional ComfyUI custom nodes for Tencent's HunyuanImage-3.0 80B multimodal model — text-to-image, Instruct text-to-image with CoT reasoning, image editing with natural language, multi-image fusion (2–5 inputs), Instruct-Distil 8-step fast inference, experimental latent/composition control, block swap, multiple quantization modes (BF16/INT8/NF4), smart memory management, and high-res generation up to 4K+.
https://github.com/EricRollei/Comfy_HunyuanImage3
git clone https://github.com/EricRollei/Comfy_HunyuanImage3
comfy node install hunyuanimage3