MiniMax H3 (Hailuo 3.0) explained
MiniMax H3 combines video generation, synchronized audio, and multimodal references in one workflow. See what is confirmed, where access exists, and what you can use in MuseFable today.
Partner source checked July 31, 2026. Provider limits can change during rollout.
MuseFable availability
Released externally; MuseFable integration is pending
MiniMax H3 is available through launch partners, but it is not yet a selectable MuseFable generation model. We do not publish a MuseFable credit price or claim an in-product workflow until the model ID and generation contract are verified end to end.
What is new in MiniMax H3?
These capabilities summarize public launch and partner materials. Exact limits can differ by endpoint, account, and rollout stage.
2K AI video
Launch access highlights 1440p output at 24 fps, aimed at sharper production-ready clips rather than low-resolution previews.
Native stereo audio
H3 can generate sound with the picture, reducing the gap between a silent visual draft and an audiovisual scene.
Omni-reference input
Text, images, video, and audio references can guide identity, movement, sound, camera direction, and style in one request.
Multiple video workflows
Public endpoints cover text-to-video, first-and-last-frame generation, and reference-led video creation for clips up to 15 seconds.
Create video in MuseFable now
You can start the same prompt-development and project workflow with a currently integrated model, then revisit H3 after its MuseFable route is verified.
Try Doubao Seedance 2.0
Generate a video from a prompt with the current MuseFable video workflow, save the result to your history, and compare future H3 output from the same brief.
MiniMax H3 frequently asked questions
What is MiniMax H3?
MiniMax H3 is MiniMax's new multimodal video model, also presented as Hailuo 3.0. It combines video generation with audio and mixed reference inputs rather than treating each medium as a separate final step.
Can I use MiniMax H3 in MuseFable now?
Not yet. This page documents the model and its rollout, while the MuseFable generator continues to show only models with a verified AIAPI route, server-side price, completed output path, and credit behavior.
Does MiniMax H3 generate audio with video?
Yes. Partner launch materials describe native stereo audio generated with the clip. Available controls and reference limits depend on the endpoint you use.
How long and what resolution are MiniMax H3 videos?
Current partner access describes 2K (1440p), 24 fps output and clips up to 15 seconds. Check the active provider documentation before building a production workflow because rollout limits can change.