MOSS-Image at native 4MP: ~25s per image on 16-24GB VRAM, huge seed variance
AxonkaiLab · reddit · 2026-10-25
A Reddit user tested MOSS-Image at native 4MP using the official ComfyUI template: 25s per image with 16-24GB VRAM on an RTX PRO 6000. It handles Krea 2-style prompts remarkably well, especially candid flash lighting, skin micro-textures, and stylized proportions. All gallery images are first-try outputs with zero cherry-picking. Key caveat: massive seed variance—every seed drastically changes composition, so seed hunting is required for specific framings.
More from Multimodal
- enGEN demo shows AI video with oil-painting cinematic look — jordanbayne · 2026-10-25
- MiniMax H3 voice reference reportedly broken: ref2va only works as direct pass-through — Beneficial_Fish_7509 · 2026-10-25
- Indie dev releases Visage-Image face T2I model; custom narrow-domain VAE beats FLUX.2 on iFID — Sherstnyov · 2026-10-25
- Stunning AI-generated street drive video looks near-photoreal end to end — Extraaltodeus · 2026-10-25
- AI animation flood incoming: Three.js slop now, Claude's telltale editing sounds — ChrisGPT · 2026-10-25
- Ecco the Dolphin Reimagined with AI Video — demondisc · 2026-10-25