World Model Index
MiniMax H3
Text- and image-to-video generation, hosted by fal with resolution-based pricing.
- Provider
- MiniMax (via fal)
- Category
- Video clips
- Published price
- $216.00 /hrGenerated video · 768p
- Resolution
- 480p / 768p / 2K / 4K options
- Frame rate
- Not published
- Access
- Public API
Capabilities
- Text to video
- Native audio
- 5–15 second clips
- Aspect ratio selection
Pricing & billing
fal’s generated-video rates convert to $180.00/hr at 480p, $216.00/hr at 768p, $468.00/hr at 2K, and $576.00/hr at 4K. The comparison uses 768p.
Billing basis: Generated video output duration, standard 768p tariff
Quality evidence
1141 Elo on AA-Video-I2V v1.1, image to video with audio. Tested entry: MiniMax H3 (768p). 95% interval: 1124–1158, from 4,009 samples.
Artificial Analysis leaderboard · checked October 15, 2026.
Video Elo measures clip preference. World consistency and interaction latency require separate tests.
Related comparisons
Notes
The published arena entry is MiniMax H3 (768p). Its score describes the tested image-to-video configuration, not every available resolution or endpoint.
Sources
Last checked October 15, 2026. Specifications and prices use provider sources; Elo links to the evaluator.
Compare models
View pricing and capabilities side by side.