Hands-on Test of Minimax H3: Multi-angle Video Generation and Long Consistency
lazyspock · reddit · 2026-08-23
A user tested Minimax H3's video generation capabilities with complex prompts focusing on multi-angle consistency, long takes, and physical logic.
Test Cases & Scores:
- Video 1 (10/10): Synchronized dual-angle (front + side) shot of a woman by the sea. Actions matched perfectly.
- Video 2 (9/10): Four-angle synchronization (front/right/left/back). Backgrounds were slightly off in some angles.
- Video 3 (8/10): Running through a hall of mirrors. Some "ghost reflections" occurred, but spatial logic held up.
- Video 4 (7/10): Simulated selfie camera rotation. The physics of the phone flip felt unnatural.
- Video 5 (6/10): Attempted to generate glass textures; the result was poor.
Overall, the model is impressive in handling complex spatiotemporal synchronization and multi-view generation, though it struggles with fine-grained physical simulations.
More from Multimodal
- AssemblyAI's Universal-3.6 Pro speech-to-text lands on OpenRouter at $0.30/hr — AssemblyAI · 2026-10-16
- FLUX 3 Image Klein lands on fal: box-drawn layouts, multi-reference placement, style transfer — bfl_ai · 2026-10-16
- SGLang ships Day-0 support for FLUX 3 Image Klein: 10 reference images, BF16/FP8/NVFP4 variants — ying11231 · 2026-10-16
- Krea 3 Scores 4.81 on CPI-Bench, Second Only to GPT Image 2.5 — testingcatalog · 2026-10-16
- Higgsfield teases Kling 4.0 rollout with a monkey-driven demo video — azed_ai · 2026-10-16
- Halcyon early-fusion image models match late-fusion with 275x less pretraining compute, open-sourced — davidwromero · 2026-10-16