MiniMax H3 Reference-to-Video by MINIMAX
minimax/h3/reference-to-videoMiniMax H3 reference-to-video: generate a video that keeps the subject from a reference image, driven by a text prompt. Supports 2K, 5-15s.
Set up the model and run your first generation.
Ctrl ⏎Examples
Runs published with this model. Load one into the playground and change whatever you like.
Little Bear straightened his clothes, getting ready to look cool.
resolution: 768Pduration: 8ratio: adaptive
Providers
Where this model can run, priced for your account. Auto routing picks the top row and moves down it if a provider fails.
| Provider | Price | Speed | Completed |
|---|---|---|---|
| fal.aiLowest price | $0.0002 /compute sec | — | — |
| Atlas Cloud | $0.100 /sec | — | — |
| Replicate | $0.080–$0.130 /sec | — | — |
| OpenRouter | — | — | — |
Speed and reliability appear once this platform has enough recent runs on a provider.
MiniMax H3 Max Image-to-Video$0.240MiniMax H3 Max Text-to-Video$0.240MiniMax H3 Fast Image-to-Video$0.220MiniMax H3 Fast Reference-to-Video$0.220MiniMax H3 Fast Text-to-Video$0.220MiniMax H3-Developer Image-to-Video$0.075MiniMax H3-Developer Reference-to-Video$0.075MiniMax H3-Developer Text-to-Video$0.075
Questions
- $0.190 per generation on ElliSekiz — the provider's own price, billed 1:1 from a prepaid USD balance with no markup and no subscription. Failed generations are free.
- fal.ai, Atlas Cloud, Replicate, OpenRouter. Requests route to the cheapest available provider automatically, or you can pin one per request or per model.
- POST to https://ellisekiz.ai/api/v1/generations with "model": "minimax-h3-reference-to-video" and your prompt, authenticated with an ElliSekiz API token. The API tab on this page has the full request for this model, including its own parameters.
- resolution, duration, ratio, prompt_expansion — each one is in the playground form on this page and in the API example.
$0.190 per generation (≈5s clip) · billed 1:1 at provider cost · compare providers