MiniMax H3 video generation launches on Japan's ai& platform from $0.0125/second
DavidBennett__ · x · 2026-10-23
Japanese inference platform ai& launched ai& video, powered by the MiniMax H3 model positioned as fast and cost-efficient.
- Generates new footage from text, photos, or existing video in as little as 12 seconds
- Pay-as-you-go pricing: H3 Fast starts at $0.0125 per second of generated video
- H3 is optimized for NVIDIA, AMD, and Tenstorrent hardware, running on domestic Japanese infrastructure
- The platform stresses data residency in Japan, in-house moderation for Japanese prompts, and that uploaded assets are never used for training
Related event: Japan's ai& Launches MiniMax H3 Video Generation(3 posts)→
More from Infra
- Jensen Huang says neocloud demand is 'off the charts' — but who funds it? — 2C_ornot2C · 2026-10-23
- $126B could carpet the entire 494-mile California high-speed rail route in $100 bills — john__allard · 2026-10-23
- TTR-Compressor: test-time regression with closed-form solutions beats pooling for sequence compression — SphereLab · 2026-10-23
- Dragonfly files CFTC comment letter urging standardized compute derivatives markets — AccBalanced · 2026-10-23
- Liquid Inference launches token Reservations: lock in fixed prices for 1-12 months with competing providers — AccBalanced · 2026-10-23
- Liquid Inference network hits 50 providers in two weeks, benchmarking every model-provider pair — AccBalanced · 2026-10-23