Leonardo adds MiniMax H3 for video with built-in soundtrack
By Steven Van ·
H3 pairs clip generation with native stereo audio and lands on Leonardo the same day as Krea, Magnific, Lovart, Pika and Runway.
Leonardo has added MiniMax H3, and the platform's framing gets at what actually sets the model apart: most video models give you a clip, H3 gives you the clip and the soundtrack, with character and voice locked in from reference material.
Built for finished, not draft, output
Leonardo is positioning H3 for commercial-grade work across ads, e-commerce, gaming, and UI, output meant to ship, not just previz. That's a meaningful claim for a model this cost-efficient: MiniMax has been marketing H3 as matching rivals like Seedance on quality while costing a fraction as much, and Leonardo's messaging leans into that as best value in its class among fully multimodal video models.
One lightweight, fully multimodal model
Under the hood, H3 (also known as Hailuo 3.0) generates 5 to 15 second clips at 2K with native stereo audio and accepts a mix of reference images, video, and audio to keep a character, product, or scene consistent across a generation. Bundling audio generation into the same pass as video removes a step Leonardo users previously had to handle in a separate tool.
Landing across the ecosystem at once
H3 arrived on Leonardo the same day it rolled out to Krea, Magnific, Lovart, Pika, and Runway, giving Leonardo users the same frontier model available across most of the category's leading platforms.
Sources: Leonardo on X.