MiniMax H3 Local Generation Fail: Severely Deformed Faces in Videos
Silver-Spot-2763 · reddit · 2026-08-06
A developer reported severe issues on Reddit when using the MiniMax H3 model for text-to-video generation. Although the model is extremely fast, using the int8 quantized pruned version in ComfyUI resulted in horrific facial deformities (like black holes for eyes and frayed edges), reminiscent of early SD 1.5 flaws.
The author noted that Wan and LTX models never exhibited such problems under the same workflow and is currently seeking community help to resolve this generation disaster.
Related event: MiniMax H3 Local Tests: Fast but Flawed in Quality and Consistency(19 posts)→
More from Multimodal
- Kling 4.0 fully available, shows markedly better motion coherence in fast-action scenes — alifcoder · 2026-10-16
- 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