MiniMax H3 API

What MiniMax H3 does

Use one API for 5–15 second H3 Max Turbo and H3 text-to-video, image-to-video, and multimodal reference-to-video. H3 reference mode accepts images, videos, and audio and covers video-to-video editing.

• Natural Motion with Frame Constraints — Upload one image to anchor the opening composition, or add a second image to define the final frame. The model fills the interval with subject motion, camera travel, and evolving light while respecting those visual anchors.
• From Fast Previews to High-Resolution Delivery — Validate motion at 480P or 768P, then move to 1080P, 2K, or 4K where the selected model supports it. The same task contract carries previews and final renders through one polling and result-handling flow.

MiniMax H3 API features

• Two Model Variants — H3 Max Turbo emphasizes detailed motion, while H3 supports output up to 4K.
• First and Last Frame Control — Provide up to two images to define the opening and ending compositions.
• 5–15 Second Output — Select duration in seconds for ad shots, product animation, and concept clips.
• Repeatable Generations — Control randomness with seed and configure safety checking and prompt expansion.

What people build with MiniMax H3

• Product Shots — Turn a product still into rotations, push-ins, and detail reveals while keeping the composition stable.
• Advertising Storyboards — Define transition intent with opening and ending frames to validate shot rhythm and visual direction quickly.
• Social Clips — Generate compact source clips for social formats, then assemble and score them in the editing workflow.
• Concept Visualization — Translate atmosphere, character action, and cinematography language into reviewable motion concepts.

Getting started with the MiniMax H3 API

The endpoint is OpenAI-compatible, so existing client code usually needs only a base URL and a model name.

• Step 1: Create an API Token — Create a token in the console for Bearer authentication.
• Step 2: Choose Model and Tier — Choose a canonical model ID, then set duration and a supported resolution.
• Step 3: Submit the Task — Send the prompt to POST /v1/videos with optional opening and ending frames.
• Step 4: Poll and Save the Result — Poll by task ID, then read metadata.url and save the completed video.

MiniMax H3 API pricing

5s H3 Max Turbo at 768P: $0.24

Why choose CCAPI for MiniMax H3

• One Task Contract — Both models share submission, polling, and result shapes, so switching variants does not require a new task pipeline.
• Published Per-Second Pricing — Per-second rates are listed by model and resolution so the application can calculate budget before submission.
• Practical Generation Controls — The API exposes duration, resolution, seed, safety checking, and prompt expansion controls needed in production.

Other models on the same endpoint

The same API key and base URL reach every model below. Moving from MiniMax H3 to any of them means changing the model name in the request body.

• Vidu API — video model on the same endpoint
• Seedance API — video model on the same endpoint
• Lyria 3 Pro API — audio model on the same endpoint
• Nano Banana API — image model on the same endpoint
• Doubao Seedream API — image model on the same endpoint
• CCAPI Upscale API — image model on the same endpoint

Frequently asked questions