MOSS-Image ComfyUI workflow runs on 8GB cards, 2 seconds per image at 4 steps
AuzBoss · reddit · 2026-10-21
The author shares a MOSS-Image ComfyUI workflow that has been doing well on Civitai, benchmarked against Qwen Image 2.1 and FLUX 3 Klein in an accompanying video.
The workflow is free, no login required, and relies on the author's AusBoss nodes (searchable in ComfyUI-Manager):
- Runs on 8GB cards with 16GB RAM; about 2 seconds per image at 4 steps on a 5090
- Includes SeedVR2 upscaling
- Can replace text on images, draw boxes to place objects, and auto-prompt for text-to-image
- The thumbnail, made with MOSS-Image itself, highlights the 5B model's text-spelling ability
Related event: MOSS-Image ComfyUI Workflow Runs on 8GB VRAM, 2s per Image(2 posts)→
More from Multimodal
- Suno Highlights Five Niche Features Most Users Don't Know About — suno · 2026-10-23
- Meshy offers Hermes users a month of Pro with API access for $1 — Teknium · 2026-10-23
- Customuse launches Smart Mesh AI with geometry-aware adaptive character clothing tool — SimplyAnnisa · 2026-10-23
- AI-generated movie poster 'The Chase' crams a wolf, a plane and muscle cars into one chase — Aggravating-Sir-115 · 2026-10-23
- SteerTTS-edge: an 11.84MB steerable TTS with unchanged WER running 19x realtime on CPU — Working_Hat5120 · 2026-10-23
- ElevenLabs' Chimera pairs Eleven v4 voice with Seedance video to make character images speak — AIwithGhotai · 2026-10-23