Open ComfyUI Node Pack: Krea 2, Z-Image, MiniMax Music 3 and Apple MLX LLMs
TimeTruth2490 · reddit · 2026-10-09
Developer lvladikov released LVNodes, a ComfyUI node and workflow collection covering Krea 2 (Turbo/Raw), Z-Image (Turbo/Base), MiniMax Music 3 for full songs with vocals, an Image2Text prompt-ready describer (Qwen3-VL), and a multi-turn LLM Chat inside ComfyUI with streaming, vision and web tools. The standout: running language models on Apple MLX straight from ComfyUI's own files — Qwen3 27B int4 at 24 tok/s, Gemma E4B at 66 tok/s (vs 1.6 via ComfyUI's native path), and MoE models like Qwen3-35B-A3B at 82 tok/s that ComfyUI otherwise can't run. All hosted on Hugging Face.
Related event: Open-Source ComfyUI Node Pack Covers Krea 2, Z-Image and Music Generation(2 posts)→
More from Multimodal
- AI image of the day: GPT-6 Image 2.5 recreates 1940s Chinese village newsreel — DeryaTR_ · 2026-10-09
- Vigglorious Studio: chunked reference frames and guide keyframes fix character drift in long AI videos — Tablaski · 2026-10-09
- Star Wars AI songs strike again: Qui-Gon gets his own 'Gin and Juice' anthem — aronchick · 2026-10-09
- MiniMax H3 zombie short film goes viral on Reddit — gabxav · 2026-10-09
- Seedance 2.5 AI video praised as a "masterpiece" in viral share — BLUECOW009 · 2026-10-09
- HeyGen Video Tops OpenRouter's Video Rankings, Hands Out Codes for 100 Free Clips Each — aziz4ai · 2026-10-09