Model Launches

MiniMax H3 on Flixly: 2K Multimodal Video with Native Audio

MiniMax H3 (Hailuo-03) is a next-generation multimodal video model — up to 2K with native stereo audio, from a text prompt or a reference image. It's now live on Flixly.

August 1, 2026

TL;DR

MiniMax H3 (Hailuo-03) is MiniMax's new multimodal video model. It generates up to 2K video with native stereo audio from a text prompt or a reference image, supports durations of 4–15 seconds and six aspect ratios, and is live in Flixly today for text-to-video and image-to-video.

MiniMax H3 — also known as Hailuo-03 — is MiniMax's next-generation multimodal video model, now live on Flixly. It generates polished short videos at up to 2K resolution with native stereo audio, working from either a written prompt or a reference image.

What makes it different

  • Native 2K + stereo audio. Video and sound come from the same pass — dialogue, music, ambience, and effects are directed in the prompt rather than added afterward.
  • Multimodal understanding. H3 interprets text and images in one context, so a reference image can anchor character or product identity while the prompt drives motion, camera, and shot structure.
  • Strong instruction following. It holds shot order, timing, camera movement, and on-screen text more reliably, and renders clearer titles and signage.

Specs

Property Value
Modes on Flixly Text-to-Video, Image-to-Video
Duration 4–15 seconds
Aspect ratios 21:9, 16:9, 4:3, 1:1, 3:4, 9:16
Resolution up to 2K
Audio native stereo, in-sync
Prompt length up to 5,000 characters

How to use it on Flixly

  1. Open Text to Video (or Image to Video if you're animating an image) in the dashboard.
  2. Pick MiniMax H3 from the model selector.
  3. For image-to-video, upload your reference image (JPG/PNG/WebP).
  4. Write a specific prompt — describe the scene, action, camera move, and any dialogue or sound you want.
  5. Set aspect ratio and duration (4–15s), then generate.

Shorter clips (4–6s) are great for fast iteration; push to 12–15s when you want the audio and pacing to breathe.

Try it

MiniMax H3 is live in Text to Video and Image to Video. Generate something, post it, and tag us — we feature the best clips in the explore gallery.

Frequently Asked Questions

What is MiniMax H3?

MiniMax H3, also called Hailuo-03, is a multimodal video generation model. It understands text and images within one creative context and produces short videos — up to 2K resolution — with native stereo audio.

What durations and aspect ratios does it support?

Durations are 4 to 15 seconds. Aspect ratios are 21:9, 16:9, 4:3, 1:1, 3:4, and 9:16. Output goes up to 2K resolution with audio generated in sync with the scene.

Does it support audio?

Yes — MiniMax H3 generates native stereo audio together with the video, so dialogue, music, and sound effects come out of the same pass with no separate TTS step.

Can I animate my own image?

Yes. Use Image to Video, upload a reference image, and add a prompt describing the motion and camera. Text to Video works from a prompt alone when you don't have a starting image.

Where can I try it on Flixly?

MiniMax H3 is live in Text to Video and Image to Video on the Flixly dashboard. Pick it from the model selector, set your duration and aspect ratio, and generate.

Tools mentioned in this post

MiniMaxMiniMax H3HailuoHailuo 03Text to VideoImage to VideoAI Video2K VideoFlixly

Ready to create with Model Launches?

Jump straight into Flixly's AI studio and try model launches with 50+ models — free to start.