MiniMax H3 | Roten AI

Scoped billing reference

Price for the stated settings.

$0.056Per second

768p · no reference video

Video is priced per second for the stated output settings.

Get API Access

Capabilities

Model capabilities.

Image → Video

Submit an image URL, check the task status and download the completed MP4 video.

Input → output

image → video

Workload candidates.

Candidates for the stated workload and billing scope.

Image → Video

Available candidate

Submit an image URL, check the task status and download the completed MP4 video.

View workload

Reference Arena scores.

Reference scores.

Reference Arena scores for MiniMax H3
Workload	Benchmark	Recorded score
Image → Video	Arena	1494

Reference benchmark data is separate from API capability tests.

Read the methodology

Implementation

Test this model.

Copy a model request, replace the API key with your own, and run.

API model ID: MiniMax-H3

Image-to-video generation. Use the sample image URL in the example and retrieve the completed video as MP4.

Copy the Python or cURL test for MiniMax H3. Replace the API-key placeholder with your own key and run.

MiniMax H3 · image to video · Uses a Roten-hosted sample image and returns the video task ID and status.

MiniMax H3 · Python

import requests

response = requests.post(
    url="https://api.roten-ai.com/v1/video/generations",
    headers={
        "Authorization": "Bearer YOUR_ROTEN_API_KEY",
    },
    json={
        "model": "MiniMax-H3",
        "prompt": "Animate the abstract shapes in this image with gentle motion while keeping the composition recognizable.",
        "duration": 5,
        "size": "768p",
        "metadata": {
            "generation_mode": "image_to_video"
        },
        "image": "https://roten-ai.com/examples/sample.png"
    },
    timeout=180,
)

print(response.json())
response.raise_for_status()

MiniMax H3 · cURL

curl --fail-with-body 'https://api.roten-ai.com/v1/video/generations' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer YOUR_ROTEN_API_KEY' \
  -d '{
  "model": "MiniMax-H3",
  "prompt": "Animate the abstract shapes in this image with gentle motion while keeping the composition recognizable.",
  "duration": 5,
  "size": "768p",
  "metadata": {
    "generation_mode": "image_to_video"
  },
  "image": "https://roten-ai.com/examples/sample.png"
}'

Python requires requests: install once with python3 -m pip install requests. cURL examples run directly in your terminal.

Video generation runs asynchronously. Use the returned task ID with GET https://api.roten-ai.com/v1/video/generations/{task_id} to check its status. When completed, download from GET https://api.roten-ai.com/v1/videos/{task_id}/content. Use the same Bearer API key for both requests.

Built for production

Running AI at scale?

Talk to us about volume pricing, model selection and migration.

Talk to Sales