Follows the prompt closely
Post-training focused on prompt adherence, so the actions, props and camera moves you write are more likely to show up.
MiniMax · H3 Max · launched August 27, 2026
MiniMax H3 Max is a speed-focused version of MiniMax's H3 video model, post-trained by fal and launched on August 27, 2026. It makes videos with synchronized dialogue, ambience, effects and music. On ViNano you can run it at 480p, 768p or 1080p for 5, 10 or 15 seconds, in six aspect ratios including 21:9.
MiniMax H3 Max needs a paid plan, and each video uses credits (see the table below). Free accounts can make videos with Seedance 1.0 Pro using their sign-up credits. ViNano is an independent platform and is not affiliated with MiniMax.
Content generated on this platform is produced by artificial intelligence models and does not represent the platform's position. [Disclaimer]. [Privacy Policy].
H3 Max builds on MiniMax H3, which MiniMax released in July 2026. According to fal's announcement, fal Research post-trained H3 for better prompt adherence, visual quality and speed. MiniMax lists it in its own API as MiniMax-H3-Max, and its Hailuo page highlights native audio.
fal reports large speed gains in its own tests; those are vendor figures, and the real wait on ViNano depends on length, resolution and queue. Compared with standard H3, Max gives up H3's 2K option and its richer video and audio references in exchange for faster output. On ViNano you can add one image as the first frame, or two for the first and last frame.
Facts on this page were checked against the official sources above in November 2026.
These are the options the generator above accepts for MiniMax H3 Max. The numbers are read from the same settings the generator uses, so they stay current when prices change.
| Resolution | 480p / 768p / 1080p |
|---|---|
| Video length | 5 / 10 / 15 seconds |
| Aspect ratio | 16:9 / 9:16 / 1:1 / 4:3 / 3:4 / 21:9 |
| Reference input | Text prompt, or up to 2 images (first and last frame) |
| Sound | Generated together with the video |
| Credits for a 5-second video | 480p 170 credits · 768p 260 credits · 1080p 575 credits |
Longer videos cost proportionally more: a 10-second video uses twice the credits of a 5-second one at the same resolution. Credits come with a paid plan or credit pack, and credits for a video that fails are returned to your balance.
H3 Max is tuned to do what the prompt says, quickly.
Post-training focused on prompt adherence, so the actions, props and camera moves you write are more likely to show up.
Dialogue, ambient sound, effects and music are generated with the picture, in sync.
Besides 16:9, 9:16, 1:1, 4:3 and 3:4, you can make cinematic 21:9 videos.
Set the opening image, or both ends of the shot, and let H3 Max animate the motion in between.
Detailed prompts play to H3 Max's strength. Try the first one in 21:9.
Ultra-wide shot of a lone rider on horseback crossing a desert at sunset, long shadows on the dunes, dust trailing behind. Slow pan right, sweeping orchestral music, warm film grain.
Two detectives sit in a parked car at night in the rain. The older one says, "He's not coming." The younger one checks her watch and replies, "Give him five minutes." Rain on the windshield, city lights, quiet tension.
A pair of white sneakers lands on a wet basketball court in slow motion, splashing water outward. The camera circles the shoes as sneaker squeaks and an upbeat beat play. High-contrast lighting.
Write what happens in the video: the subject, the action, the camera move and the mood. One clear scene per prompt works best.
Upload a photo to animate it. Where MiniMax H3 Max accepts a last frame, add a second image to set where the video ends.
Choose 16:9 for YouTube, 9:16 for TikTok and Reels, or 1:1 for feeds. The button shows the credits before you start.
Most videos are ready in one to three minutes. Download the MP4, or find it later in your video history.
Wide 21:9 scenes for trailers, music videos and film-style intros.
Ads where several specific details need to appear just as written.
Short dialogues with voices and room sound generated together.
Turn a concept image into motion with a first frame, or plan a reveal with first and last frames.
H3 Max is the faster MiniMax model. Here is how it lines up with similar models on ViNano.
| Model | Resolution | Length | Sound | 5-second video from |
|---|---|---|---|---|
| MiniMax H3 Max | 480p / 768p / 1080p | 5 / 10 / 15 seconds | Yes | 170 credits |
| MiniMax H3 | 768p / 2K | 5 / 10 / 15 seconds | Yes | 200 credits |
| Kling 3.0 | 720p / 1080p | 5 / 10 / 15 seconds | Yes | 455 credits |
| Wan 3.0 Prime | 480p / 720p / 1080p | 5 / 10 / 15 seconds | Yes | 230 credits |
No. MiniMax H3 Max is available on paid plans, and each video uses credits; the table above lists the cost for every resolution. Free accounts can make AI videos with Seedance 1.0 Pro using their sign-up credits.
H3 is MiniMax's quality-first model, with 2K output and image, video and audio references. H3 Max is post-trained by fal for speed and prompt adherence and drops the 2K option. Both generate sound.
It is based on MiniMax's H3 model and was post-trained and launched by fal on August 27, 2026. MiniMax also offers it in its own API and on Hailuo.
fal and MiniMax both advertise it as fast, and fal reports renders in seconds on its own servers. Those are vendor figures; on ViNano most videos are ready in one to three minutes, including time in the queue.
16:9, 9:16, 1:1, 4:3, 3:4 and 21:9. If you start from an image, choose the ratio closest to it for the best framing.
We haven't found an open-weight release of H3 Max; it is offered through MiniMax's and fal's APIs. ViNano runs it in the cloud, so you don't need your own GPU.