MachGen pushes MiniMax H3 past its 15s cap with 30-second continuous video
MiniMax_AI · x · 2026-10-02
MiniMax reshared third-party inference platform MachGen's announcement: MachGen now offers 16-30s continuous generation for MiniMax H3, breaking past the model's native 15-second cap — a single uninterrupted sequence, not stitched clips. Available now in open beta on UI and API. MachGen claims major speedups (up to 15x on H3) for open image/video models like MiniMax H3, LTX 2.3, HiDream O1 and Flux 2 Dev via an optimized inference platform.
More from Infra
- US egocentric cleaning data (100s of hrs/week capacity) can't even sell at break-even pricing — paigeinsf · 2026-10-02
- Buyers now fill out export control declarations when purchasing RTX 5090s in stores — blelbach · 2026-10-02
- Redditor builds fully local LLM-powered radio site on two DGX Sparks and a 5090 — jwhh91 · 2026-10-02
- VC quip: many neoclouds are closer to 95% than five nines of reliability — saranormous · 2026-10-02
- Report: lenders demand up to 25% collateral from Nvidia as GPU-backed loans wobble — GaryMarcus · 2026-10-02
- Microsoft Backs Snowflake-Led Effort to Standardize Business Metrics for AI — xiaosun86 · 2026-10-02