←Back to NewsAI News/ImagemodelImageDevelopmentMitsuba Squeezes a 27B Vision Model Into 7.3 GB for ComfyUIA 7.3GB ternary-quantized Qwen3.8-27B vision model tuned specifically for writing ComfyUI image and video generation prompts from reference images.SourceAlphaSignalPublishedOct 1, 2026, 12:17 PMAuthorAlphaSignal NewsroomRead1 min readA 7.3GB ternary-quantized Qwen3.8-27B vision model tuned specifically for writing ComfyUI image and video generation prompts from reference images.Reporting is indexed from AlphaSignal. Rights remain with the original publisher and cited sources.Read original report ↗Next readsGoogle Research · newsGoogle's Diffusion Controller Beats LoRA and Wins 90% of Image Quality TestsRunway · newsRunway Adds Seedream 5.0 Layers to Split Flat Images Into Editable PartsMidjourney · newsMidjourney's V8 Drops Seamless Tiles, Smarter Edits, and Live Style Previews
