MiniMax releases MiniMax-H3, an open-weights 33B image-to-video model — community tooling lands within a day

MiniMax published MiniMax-H3, a 33-billion-parameter open-weights image-and-text-to-video model, on Hugging Face in early August 2026. By its August 6 update it topped the trending chart, with ComfyUI packaging and community turbo-LoRA variants appearing within a day.

What it means Open-weights video generation at this scale moves a capability that was API-only last year onto self-hosted rigs — and the day-one tooling shows real demand.