MiniMax Hailuo H3 Lite Image to Video

MiniMax Hailuo H3 Lite Image to Video

MiniMax Hailuo H3 Lite Image to Video turns a single still image into a short, coherent video clip. It is the lightweight, low-cost tier of MiniMax's omni-modal H3 family, built for exploration and high-volume batch generation at 480P or 768P.

What is MiniMax Hailuo H3 Lite Image to Video?

H3 Lite Image to Video animates a supplied frame. Give it a first frame, a last frame, or both, add a prompt describing the motion, and it generates a 4-15 second clip where your still comes to life. Sharing the H3 lineage, it follows instructions closely: the subject, wardrobe, setting and specific action you describe appear in the result, and fabric, hair, liquids and machinery move with physically plausible weight. The output aspect ratio follows the image you supply.

Key Features

• •Image-to-video from a first frame, a last frame, or interpolation between both
• •480P or 768P output; 4 to 15 second clips
• •Strong instruction following and accurate text and brand rendering from the H3 family
• •Prompts up to 7000 characters for detailed motion direction
• •Output aspect ratio inherited from the input image

Best Use Cases

Lite is the throughput tier: fast clips for storyboarding, social-first content, and batch animation of product, fashion and advertising stills. In testing, a single still of a runway model in a crimson silk gown animated into a continuous shot where she strides forward and the long train whips across the floor - bringing a product or character pose to life. Use it to test shot ideas before committing to a higher-fidelity tier such as base H3 or H3 Max.

Prompt Tips and Output Quality

Compose the input so the thing that should move is already in frame, then write the prompt as the motion, not the scene. One clear, verb-driven action reads most cleanly. Keep legible text and dense crowds modest; the Lite tier trades fine detail for speed. 768P is the top quality here (no 2K); drop to 480P for the fastest batches.

How to use MiniMax Hailuo H3 Lite Image to Video

MiniMax Hailuo H3 Lite Image to Video animates a still image into a short clip. The workflow is simple: upload a frame, describe the motion, pick a duration and resolution, and generate.

Choosing your frames

Supply a first_frame_image to animate forward from a single still - the most common use. Add a last_frame_image to define a target end state, and the model interpolates between the two frames; you can also send a last frame on its own. Pick an input where something can visibly move: billowing fabric, hair, steam, flowing liquid, a vehicle, a dancer. The output aspect ratio follows the image you supply, so frame your still at the ratio you want the video to be.

Writing the prompt

Because the subject already lives in the image, write the prompt as the motion rather than re-describing the scene. Name one clear, verb-driven action ("she strides forward and the train whips across the floor") and the camera behaviour if it matters. Prompts can run up to 7000 characters, but a single specific action reads most cleanly on the Lite tier.

Recommended Settings

• •resolution: 768P for the best quality this tier offers; 480P for the fastest batch runs. There is no 2K on Lite.
• •duration: 6 seconds is a solid default that fits one complete action; the range is 4 to 15 seconds.
• •first_frame_image: a clean, well-composed still where the motion subject is already present.

In testing, a 768P, 6-second run from a single runway still produced a clean 1344x768 clip with physically coherent fabric and body motion animated straight from the frame.

Tips

Keep tiny legible text and dense crowd motion modest - the Lite tier favours speed over fine detail. Draft and iterate here, then re-run the winning shot on base H3 or H3 Max for a higher-fidelity final.

Examples to try

Each example below ran end to end on this page at the settings shown, animating a supplied still (768P / 6s for the default, 480P / 5s for the two newer cards); the output aspect ratio follows the input frame every time.

• •Motion from a single still (default). Prompt: "The runway model strides forward down the glossy black catwalk; the crimson silk gown billows and snaps with each step and the long train whips across the polished floor as she pivots hard at the end. Overhead spotlights rake the runway." The posed model begins striding in the first second and the silk gown and train physically billow and sweep. Supply a still where the fabric or hair can already move, and write the prompt as the motion.
• •First- and last-frame interpolation. Prompt: "The flat-folded paper crane lifts off the walnut table, its creased wings spreading wide and its neck straightening until it stands upright, the sharp folds catching the light." With both a first frame and a last frame supplied, the clip morphs from the folded crane to the upright one. Keep the two stills at the same framing and lighting for a clean transition.
• •Fluid pour motion. Prompt: "Dark chocolate pours in a glossy ribbon from the tilted copper jug and spreads across the top of the tart, rippling outward and settling as the jug lifts away." The chocolate streams from the jug and spreads over the tart, matching the action named in the prompt. Name one action and compose the still so whatever should move is already in frame.

More to try: "steam curls up from the fresh bowl of ramen," "the sports car's wheels spin as it pulls away from the kerb," "her hair and scarf lift in a sudden gust."

Does it need an image?

Yes - supply a first frame, a last frame, or both. For text-only or reference-driven generation, use the H3 Lite text-to-video or reference-to-video siblings.

What resolutions and lengths are supported?

480P and 768P (no 2K), and 4 to 15 seconds.

What aspect ratio is the output?

It follows the input image you provide.

How do first and last frames work together?

Supply both to interpolate between them; supply only a first frame to animate forward from it.