MOSS-Image samples impress Redditors as Fudan team's new image model surfaces
Only-Baseball-3129 · reddit · 2026-10-17
A Reddit user shares sample images from MOSS-Image, saying the quality blew them away. MOSS is the Fudan University team's LLM project; its image generation branch is drawing attention, though the post offers few details or benchmarks.
More from Multimodal
- AI-generated Star Wars stills of Leia and Vader show cinematic-level image quality — azed_ai · 2026-10-17
- Procedural meta-human in Three.js adds skin oiliness, sweating and finer hair control — jasonkneen · 2026-10-17
- Midjourney sref combo 3540389609 + 155206446 yields torn-collage Nike fashion visuals — sergeantsref · 2026-10-17
- Moody foggy city street AI image: motion blur and neon glow showcase — LudovicCreator · 2026-10-17
- Kling's new model generates 30-second clips in one shot, dubbed 'Sora killer' — aitrendz_xyz · 2026-10-17
- Creator shares full AI music video pipeline: Claude for lyrics, Suno for song, local LTX for video — elonmusk · 2026-10-17