MiniMax open-sourced its H3 video generation model on August 29, alongside the launch of H3 Max, a production-grade version co-designed with a custom inference stack by fal. In human preference tests, H3 Max ranked first on overall quality, prompt following, and aesthetics at approximately 35 times the official H3 throughput. The open-source H3 model generates a 15-second 768p clip in about 13 seconds on a single GPU—roughly 14 times faster than prior methods. This release marks a significant step forward in accessible AI video generation technology.
WeChat OA
China Supply Info
Scan to Follow
Procurement
Post Product
- Post Supply
- Manage Supplies
