MiniMax H3 Max Turbo Image to Video API

MiniMax H3 Max Turbo Image to Video

MiniMax H3 Max Turbo Image to Video is built for high-throughput visual production, retaining flagship image quality while reducing latency and cost. Use a single first frame or paired start and end frames for precise transitions, with native synchronized sound at 480p, 768p, or 1080p. It preserves source-image textures while generating physically plausible camera work, supporting efficient batch production of character animation and promotional assets.

Why choose the MiniMax H3 Max Turbo Image to Video API?

• Precise transitions between start and end framesGenerate from a single opening frame or paired start and end frames. The end frame anchors the final camera position and target pose.

• Faithful subject and texture preservationClosely anchors the source character's face, clothing textures, and lighting to reduce feature drift and distortion during large movements.

• Native synchronized stereo audioAutomatically synthesizes ambient and action sounds to match the visuals and motion rhythm, delivering a complete audiovisual clip out of the box.

• Twice the throughput for faster productionPost-training distillation reduces inference sampling overhead, generating at about 2 times the speed of the base flagship model and accelerating batch production.

• Half the rates for better valueEvery resolution tier costs half as much as the base H3 Max, with pricing from 5 credits per second ($0.025/second) to reduce production costs at scale.

Parameters

How to Use

1. 01

   Upload the opening imageProvide a starting reference image with a clear subject and composition, using an HTTP(S) URL or Base64.

2. 02

   Add an optional end frameTo specify an ending pose or a precise transition, append a second image as the transition's endpoint.

3. 03

   Write a motion promptDescribe the desired actions, camera path, and atmosphere. Avoid instructions that conflict with features in the reference image.

4. 04

   Configure output settings and resolutionChoose a whole-second duration from 5–15 seconds and 480p, 768p, or 1080p output to match your production needs.

5. 05

   Submit the task and retrieve the rendered videoSubmit a generation request to receive a task_id, then poll the status endpoint or use a callback URL to receive the video with synchronized audio.

Usage and billing

Cost depends on the requested duration parameter and selected resolution tier, starting at 5 credits per second. Rates are fixed for each tier, with transparent per-second billing and a live cost estimate in the dashboard before submission.

Best Use Cases

• Portrait and character animationBring character illustrations, concept art, or photos to life with stable facial identity and realistic body movement.

• E-commerce showcases and product camera workGenerate HD showcase videos with orbiting camera views and shifting light from a single still product image.

• Transitions between film keyframesConnect two storyboard keyframes using start and end frame control to generate smooth, physically plausible transitions.

• Frequent social media video productionQuickly turn batches of still visual assets into short clips with sound effects for efficient promotional coverage across platforms.

Pro Tips

• Keep the first image's subject clear and sharp. An aspect ratio from 0.4 to 2.5 is recommended; the video automatically follows the source image's aspect ratio.
• Focus the prompt on movement developing from the reference image. Avoid repeating static details already visible in the image to keep the model's attention focused.
• When using start and end frames, keep the subject's identity, lighting, and style broadly consistent. Large differences can produce abrupt distortions.
• For a specific pan, push-in, or pullback, put camera movement before the subject's action, such as "the camera pushes in smoothly as the character turns and smiles."
• Avoid instructions that strongly conflict with the source image's colors or clothing to preserve the subject's identity.

Notes

• image_urls accepts 1–2 images: the first is the opening frame and the second is an optional end frame. Supports HTTP(S) URLs or Base64 encoding.
• In image-to-video mode, the output automatically inherits the input image's aspect ratio. No separate aspect_ratio parameter is needed.
• Successful submission immediately returns an asynchronous task_id. Poll the status endpoint or configure callback_url to receive the result.
• 1080p output uses latent refinement upsampling from 768p. For everyday batch production, the native 768p tier is recommended for better speed.
• Generated videos include a natively synthesized two-channel stereo soundtrack by default. No additional audio fields are required.