MiniMax H3 Max is a text-to-video model on the fal.ai platform, with the API path minimax/h3-max/text-to-video.
Purpose: To convert text prompts into video content, suitable for scenarios that require generating video clips directly from text descriptions.
Key capabilities: According to the information, it is a post-trained variant of MiniMax H3, tuned for stronger prompt adherence and better aesthetics, and co-optimized with fal's custom inference stack to achieve higher throughput without sacrificing output quality.
Usage: Called via the fal.ai platform's minimax/h3-max/text-to-video API, users can customize inputs for more control.
Resolution and billing: The model offers three resolution tiers: 480p, 768p, and 1080p. Billing is calculated by the number of seconds of generated video:
- Promotional period (the information mentions a limited-time 50% discount): 480p is $0.03 per second, 768p is $0.048 per second, and 1080p is $0.096 per second; after the discount ends, 480p is $0.05 per second and 768p is $0.08 per second.
- Another promotional period (the information mentions a limited-time 40% discount): after the discount ends, 480p is $0.05 per second, 768p is $0.08 per second, and 1080p is $0.16 per second.
Applicable scenarios and limitations: The information does not provide specific applicable scenario descriptions, input parameter details, maximum video duration, concurrency limits, or regional availability.
