MiniMax H3 in Action: Mixing Characters with Reference Images
Glittering_Tie_3110 · reddit · 2026-08-08
The author shares practical experiments using the MiniMax H3 model, focusing on leveraging reference images to improve character likeness and attempting to mix characters from different media franchises.
The findings suggest that matching the reference image's aspect ratio to the target video and strictly following the official prompt writing guide significantly improves generation quality, even at a lower resolution of 0.4 MP.
More from Multimodal
- MJev-Spatial: Single-Pass Classification Beats Autoregressive VLMs on Spatial Reasoning — mindlab-research · 2026-10-16
- H-DDCM: Hadamard Codebooks Make Diffusion-Based Image and Video Compression Scalable — Technion · 2026-10-16
- ComfyUI v0.41.0 ships with Flux3 image model and MiniMax-H3 LightVAE support — Fcking_Chuck · 2026-10-16
- AI-generated short film 'The Dragon Rider' glides cinematically above the clouds — jijo6686 · 2026-10-16
- Realistic candid Midjourney portrait prompt: natural skin, lived-in bedroom light — tisch_eins · 2026-10-16
- 1986 to 2026 in one session: outpaint, modernize and 4x upscale, session file included — juanpablogc · 2026-10-16