How is MiniMax H3 Max different from MiniMax H3?

Max is a post-trained version of H3, optimized for generation speed. Resolution: Max outputs 480P or 768P, H3 outputs 768P or 2K. Length: Max takes 5–15 seconds, H3 takes 4–15. Parameters: Max adds prompt_expansion_mode. Billing: Max charges reference video at its own rate and only the first two reference images are free; H3 charges reference video at the output rate and the first five reference images are free. Inputs are the same.

How do I pass prompt_expansion_mode?

Add a prompt_expansion_mode field to the POST /v1/videos body with the value "disabled", "balanced" or "quality". MiniMax’s native extra.prompt_expansion_mode is accepted too. Leaving it out uses the upstream default, balanced. Only MiniMax H3 Max accepts this parameter; other models return a 400.

How do I set resolution and length?

POST /v1/videos with model minimax/minimax-h3-max, size "480P" or "768P", seconds 5–15 and an optional ratio such as "9:16". Sending 4 seconds returns a 400 before any job is created. Leaving them out gives 768P, 5 seconds and 16:9.

MiniMax H3 Max

Save 50%Official

minimax/minimax-h3-max

MiniMax H3 Max — a faster, post-trained version of H3: 5–15 sec at 480P or 768P, text-to-video, first/last frame and multi-asset reference input.

Get an API key View code sample

Live model playground

Try MiniMax H3 Max here

Tune the request, see the current price ceiling, and inspect the real response without leaving this model page.

asyncvideo

1Configure input

Prompt

0 / 20,000

First frame

Last frame

Reference images

Maximum estimated cost—

2Live result

Your result will appear here

Enter an input on the left. The same RouteMux Playground execution and billing path powers this embedded test.

Input modalities: TextImage
Output modalities: Video
Capabilities: Video generation · Text-to-video · Image-to-video · Reference voices · Reference images · Reference videos · First and last frame
Upstream release: 2026-10-26
RouteMux name: minimax/minimax-h3-max
Official ID: MiniMax-H3-Max

What is MiniMax H3 Max

MiniMax H3 Max is a video generation model post-trained on MiniMax H3 and optimized for faster generation. MiniMax’s documentation says it generates faster than MiniMax H3 and outputs the two mainstream resolutions, 480P and 768P.

Inputs work as on H3: from a prompt alone (text-to-video), from a first frame or a first and last frame (image-to-video), or from references: up to nine reference images, three reference videos and three reference audio clips (videos and audio each 2–15 seconds per clip, 15 seconds in total). Frame images and references cannot be mixed in one request, and a request takes at most 12 input files. Duration is any whole number of seconds from 5 to 15 (4 seconds is not supported), and the aspect ratio can be 16:9, 9:16, 1:1, 4:3, 3:4 or 21:9. Prompts can run to 7,000 characters.

Max adds one optional parameter, prompt_expansion_mode, which controls how the upstream expands your prompt: disabled turns expansion off, balanced (the default) expands moderately, and quality favors output quality. Choose disabled when you want the video to stay close to your original prompt.

Pricing is per output second, with one rate for 480P and another for 768P. Reference video adds a charge per input second at its own rate, which differs from the output rate (see pricing); the first two reference images are free and each one after that is charged; reference audio is free. It uses the OpenAI Sora protocol and is asynchronous: the request returns a job, and you poll it for the finished video.

Before you choose

How is MiniMax H3 Max different from MiniMax H3?: Max is a post-trained version of H3, optimized for generation speed. Resolution: Max outputs 480P or 768P, H3 outputs 768P or 2K. Length: Max takes 5–15 seconds, H3 takes 4–15. Parameters: Max adds prompt_expansion_mode. Billing: Max charges reference video at its own rate and only the first two reference images are free; H3 charges reference video at the output rate and the first five reference images are free. Inputs are the same.
How do I pass prompt_expansion_mode?: Add a prompt_expansion_mode field to the POST /v1/videos body with the value "disabled", "balanced" or "quality". MiniMax’s native extra.prompt_expansion_mode is accepted too. Leaving it out uses the upstream default, balanced. Only MiniMax H3 Max accepts this parameter; other models return a 400.
How do I set resolution and length?: POST /v1/videos with model minimax/minimax-h3-max, size "480P" or "768P", seconds 5–15 and an optional ratio such as "9:16". Sending 4 seconds returns a 400 before any job is created. Leaving them out gives 768P, 5 seconds and 16:9.

MiniMax H3 Max API pricing

OutputOfficial $0.05 /sec

$0.025/secSave 50%

480P$0.025 /secOfficial ~~$0.05~~
768P$0.04 /secOfficial ~~$0.08~~
Reference video · 480P$0.02765 /secOfficial ~~$0.0553~~
Reference video · 768P$0.0715 /secOfficial ~~$0.143~~
Reference images (first 2 free, then each)$0.037Official ~~$0.074~~

Key capabilities

Faster generation: Post-trained on H3 and optimized for speed.
5–15 seconds: Any whole-second length; 4 seconds is not supported.
480P or 768P: Two output resolutions, each with its own per-second rate.
Optional prompt expansion: prompt_expansion_mode: disabled, balanced (default) or quality.
50% below official pricing: Billed per output second; both per-second rates are in the pricing section above.

Limited time10% extra Credits on your first top-up — up to 200 bonus Credits

Applies to your first top-up within 72 hours of signing up. Minimum $20, once per account.

Interface and protocols

OpenAI Sora

Use any of these IDs to call this model via the API.

RouteMux nameminimax/minimax-h3-max

Official IDMiniMax-H3-Max

Replace the ROUTEMUX_KEY placeholder with your API key. Create one →

curl https://api.routemux.com/v1/videos \
  -H "Authorization: Bearer <ROUTEMUX_KEY>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "minimax/minimax-h3-max",
    "prompt": "A quiet product shot of a glass observatory at dusk",
    "seconds": 5,
    "size": "1280x720"
  }'

base_url: https://api.routemux.com/v1

Performance (last 30 days)

1 requests · last 30 days

Total usage—

P50 first-token latencyInsufficient data (1 requests)

Avg throughputInsufficient data (1 requests)

Success rate—

Uptime · 24h

—

Success rate · 60d · from real requests

MiniMax H3 Max FAQ

›What is MiniMax H3 Max?

MiniMax H3 Max — a faster, post-trained version of H3: 5–15 sec at 480P or 768P, text-to-video, first/last frame and multi-asset reference input.

›How much does the MiniMax H3 Max API cost?

Through RouteMux, MiniMax H3 Max is billed per second, by resolution: 480P $0.025, 768P $0.04. Each is below the official list price.

›How do I call MiniMax H3 Max via API?

MiniMax H3 Max is OpenAI-compatible: point your base URL at RouteMux and set the model field to minimax/minimax-h3-max — no code changes needed.

›Which input and output modalities does MiniMax H3 Max support?

MiniMax H3 Max accepts text, image as input and produces video as output.

›When was MiniMax H3 Max released?

MiniMax H3 Max was released on October 26, 2026.

Sources

Reviewed by the RouteMux team on 2026-10-26

View as Markdown