MiniMax H3-Image handles 5 object references in one scene, mostly holds layout
Horror-Egg-6942 · reddit · 2026-11-12
A Reddit user tested the open-weight MiniMax H3-Image with 5 separate object reference images (rotary phone, glass cube, apple, staircase, chrome sphere), explicitly labeling Reference 1–5 to lock layout. The model supports up to 16 references (64 MiB each, 128–8192px). Results: all objects stayed identifiable with consistent lighting; the intended spiral arrangement collapsed into a simple vertical stack.
More from Multimodal
- MiniMax H3-Image merges up to 16 reference images into one concept scene — mhdfaran · 2026-11-12
- MOSS Image with 4-step LoRA runs on-device on iPhone: ~30s for 1024x1024 — woadwarrior · 2026-11-12
- Massive Gemini prompt specs a deterministic 3-minute looping violin concert film — CurieuxExplorer · 2026-11-12
- Seedream 5.5 launches on fal with multi-turn editing, in-image text edits, and layer separation — Jamez_Chi · 2026-11-12
- Runway adds 3D: text-to-3D via Tripo H3.1 with video workflow integration — runwayml · 2026-11-12
- AI artwork "Nox in Oculis" painted entirely in Python with Fae × Fable 5.2 — amplifiedamp · 2026-11-12