MiniMax H3 API: price per second, request and code

What a MiniMax H3 clip costs

Credits per clip from text, by resolution and length (5, 8 or 10 seconds). Dollars are at the headline credit price; plans with more credits pay less per credit.

Computed from the quote the API charges with, as of 15 Nov 2026. The shortest, cheapest run is 50 credits; the longest at 2K is 240.

One request, then wait for the job

Video takes one to several minutes, so POST /api/v1/videos charges the run and answers at once with a job in queued. Read GET /api/v1/jobs/{id}?wait=50: it holds the line for up to 50 seconds and returns early the moment the job is final. The body below is the cheapest valid MiniMax H3 request, 5 seconds at 480P, which costs 50 credits.

curl -X POST https://kinoseed.app/api/v1/videos \
  -H "Authorization: Bearer ks_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "a lighthouse at dusk, waves breaking, slow aerial orbit",
    "model": "minimax-h3",
    "duration": 5,
    "aspect": "16:9",
    "resolution": "480P"
  }'

Send the same body to Quote first if you want the number before you spend anything; it charges nothing. Fields and errors are on Videos and Jobs.

The same run from an agent over MCP

An agent connected to the MCP server sends the same arguments to the generate_video tool. It returns the job at once, and the agent follows it with get_job until the status is completed or failed. Results are URLs, never inline video, so a clip does not fill the context window. Setup for Claude Code, Cursor and others is in the MCP server for video generation guide.

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "generate_video",
    "arguments": {
      "prompt": "a lighthouse at dusk, waves breaking, slow aerial orbit",
      "model": "minimax-h3",
      "duration": 5,
      "aspect": "16:9",
      "resolution": "480P"
    }
  }
}

What MiniMax H3 accepts

Every field is optional except prompt and model; anything outside the model's range is snapped to the nearest value it allows, and the quote follows the snapped value.

• duration: 5, 8 or 10 seconds.
• aspect: 16:9, 9:16 or 1:1.
• resolution: 480P, 768P or 2K.
• startImageUrl: a first frame to animate. It does not change the price.
• endImageUrl: a last frame, so the clip lands on a set picture.
• Sound: Clips come back silent on Kinoseed; there is no sound option for this model.

Input images must be public https URLs or URLs returned by POST /files. The API refuses private addresses, localhost and plain http before anything fetches them.

What moves the price

Resolution and length. At 5 seconds, MiniMax H3 is 50 credits at 480P and 120 at 2K. Doubling the length roughly doubles the price, since the provider bills per second of output.

For drafts, a cheaper model often does the job: a 5 second Wan 3.0 clip at 480p is 50 credits. Test the prompt there, then run the keeper on MiniMax H3. Every model's rate per second is side by side in AI video API price per second.

When to pick MiniMax H3

Pick MiniMax H3 for big, energetic motion from a prompt or a photo, and when the clip has to end on a set picture: it takes a start frame and an end frame, and runs up to 2K.

For a clip with sound, use Seedance 2.5 or Grok Imagine. Studio users get MiniMax H3 in the Video studio; the MiniMax H3 model page has its prices and specs, and how to use MiniMax H3 covers prompts.

Failures, retries and limits

A run that fails is refunded automatically, and the job says so, so a provider error never costs credits. Identical requests sent close together return the same job instead of charging twice, and an Idempotency-Key header makes that explicit for retries from your own code.

An account runs up to 5 generations at once; a sixth request gets a 429 too_many_running error with a Retry-After header, before anything is charged. Results carry no watermark. Rate limits and error codes are on Errors and Rate limits.