MiniMax H3
Capabilities & pricing
MiniMax H3 is designed for scenes with lively character movement, expressive faces and atmospheric surroundings.
- Base rate
- from 22.5 tokens per second
- Category
- Video
- Access
- MCP + website
- Payment
- Syntasy tokens
What can you do with MiniMax H3?
Character through movement
Describe a gesture, expression and reaction. These details define a scene with a clear emotion rather than a vague request for an attractive video.
A deliberate short scene
Set the lighting, depth and camera movement. Explore portrait shots, advertising sketches and visual tests before filming.
A scene from a written brief
Start without a prepared first frame. Describe the setting, action and camera movement, then choose generation settings.
Animate a source image
The image sets the scene's starting point. Focus the prompt on movement, the action's progression and camera behavior relative to that frame.
Where it fits
Character stories, emotional scenes, advertising sketches and visual concepts.
How to get a useful result
Focus a short scene on one emotion and one action. Check this version's modes for supported source files.
How much does it cost?
Current tariffs from the Syntasy catalog. Choose a mode and settings; check the final estimate before starting.
| Mode | Settings | Syntasy tokens |
|---|---|---|
| Text to video | 768p · 21:9 · 4–15 s · Text input | 22.5 tokens per second |
| Text to video | 768p · 16:9 · 4–15 s · Text input | 22.5 tokens per second |
| Text to video | 768p · 4:3 · 4–15 s · Text input | 22.5 tokens per second |
| Text to video | 768p · 1:1 · 4–15 s · Text input | 22.5 tokens per second |
| Text to video | 768p · 3:4 · 4–15 s · Text input | 22.5 tokens per second |
| Text to video | 768p · 9:16 · 4–15 s · Text input | 22.5 tokens per second |
| Image to video | 768p · 4–15 s · With image | 22.5 tokens per second |
| Reference to video | 768p · 21:9 · 4–15 s · With reference | 22.5 tokens per second |
| Reference to video | 768p · 16:9 · 4–15 s · With reference | 22.5 tokens per second |
| Reference to video | 768p · 4:3 · 4–15 s · With reference | 22.5 tokens per second |
| Reference to video | 768p · 1:1 · 4–15 s · With reference | 22.5 tokens per second |
| Reference to video | 768p · 3:4 · 4–15 s · With reference | 22.5 tokens per second |
| Reference to video | 768p · 9:16 · 4–15 s · With reference | 22.5 tokens per second |
| Reference to video | 768p · adaptive · 4–15 s · With reference | 22.5 tokens per second |
| Text to video | 2k · 21:9 · 4–15 s · Text input | 36.5 tokens per second |
| Text to video | 2k · 16:9 · 4–15 s · Text input | 36.5 tokens per second |
| Text to video | 2k · 4:3 · 4–15 s · Text input | 36.5 tokens per second |
| Text to video | 2k · 1:1 · 4–15 s · Text input | 36.5 tokens per second |
| Text to video | 2k · 3:4 · 4–15 s · Text input | 36.5 tokens per second |
| Text to video | 2k · 9:16 · 4–15 s · Text input | 36.5 tokens per second |
| Image to video | 2k · 4–15 s · With image | 36.5 tokens per second |
| Reference to video | 2k · 21:9 · 4–15 s · With reference | 36.5 tokens per second |
| Reference to video | 2k · 16:9 · 4–15 s · With reference | 36.5 tokens per second |
| Reference to video | 2k · 4:3 · 4–15 s · With reference | 36.5 tokens per second |
| Reference to video | 2k · 1:1 · 4–15 s · With reference | 36.5 tokens per second |
| Reference to video | 2k · 3:4 · 4–15 s · With reference | 36.5 tokens per second |
| Reference to video | 2k · 9:16 · 4–15 s · With reference | 36.5 tokens per second |
| Reference to video | 2k · adaptive · 4–15 s · With reference | 36.5 tokens per second |
This is the rate for the specified billing unit. Duration, output count and input length may change the total.
Available modes
Text to video
Text input
Image to video
With image
Reference to video
With reference
Resolutions: 768P, 2K
Aspect ratios: 21:9, 16:9, 4:3, 1:1, 3:4, 9:16, adaptive
Use MiniMax H3 via MCP
Connect Syntasy in your AI client and authorize access. Ask it to find this model, calculate the price and show the settings before generation.
- Open Access & tokens in the MCP section and create an MCP token.
- Add the server URL to your client:
/api/mcp - Ask: “Find MiniMax H3. Show its modes and estimate the cost before starting.”
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "syntasy_list_models",
"arguments": {
"kind": "video",
"query": "MiniMax H3"
}
}
}The example discovers models and does not start a paid generation. Replace YOUR_MCP_TOKEN with your own token.
Frequently asked questions
What are Syntasy tokens?
They are the shared balance used to pay for models in Syntasy. Token pack prices are available on the pricing page. View token packs
Why can the final cost differ from the base rate?
A per-second tariff is multiplied by duration, an image tariff by output count, and a character tariff by text length. Check the estimate on the website or use syntasy_estimate_generation via MCP for your chosen settings.
