Animate images into 1080p video with the MiniMax H3 Max API on Muapi. Start/end frame control, pay per generation. MiniMax H3 Max Image to Video animates a source image with a motion prompt at up to 1080p through the Muapi API.
MiniMax H3 Max Image to Video animates a source image with a motion prompt at up to 1080p through the Muapi API.
Tentang model ini
MiniMax H3 Max Image to Video animates a source image into a moving video clip at up to 1080p resolution, guided by a motion prompt. Supply a starting frame with image_url, an optional end_image_url to anchor the final frame, and describe the camera movement and action you want in between. Durations run from 5 to 15 seconds across six aspect ratios, with an optional prompt-expansion pass for richer motion detail. It follows Muapi's standard asynchronous request/poll/webhook pattern. For pure text-to-video generation without a source image, see MiniMax H3 Max Text to Video; to blend multiple image, video, and audio references, see MiniMax H3 Max Reference to Video.
Analisis biaya
| Penyedia | Biaya | Catatan |
|---|---|---|
| muapiapp | From $0.0325/sec (480p) up to $0.104/sec (1080p) | Launch pricing through Sep 30, 2026 (UTC): From $0.0325/sec (480p) up to $0.104/sec (1080p). From Oct 1, 2026 the standard rate applies: $0.065–$0.208/sec by resolution. Pay-per-generation, no subscription. |
| Replicate | Tidak tersedia | MiniMax H3 Max is not listed on Replicate. |
Launch pricing through Sep 30, 2026 (UTC): From $0.0325/sec (480p) up to $0.104/sec (1080p). From Oct 1, 2026 the standard rate applies: $0.065–$0.208/sec by resolution. Pay-per-generation, no subscription.
MiniMax H3 Max is not listed on Replicate.
** Harga pesaing diperkirakan berdasarkan arsitektur model dan tingkat penggunaan yang serupa.
Skema konfigurasi
| Parameter | Jenis | Keterangan | Bawaan |
|---|---|---|---|
| cepat | string | , . | The camera slowly pushes in as the subject turns toward the light and fabric moves in a soft breeze. |
| URL gambar | string | Starting frame image. The output canvas follows this image. | undefined |
| URL Gambar Akhir | string | Optional final frame image. Can be used alone for end-keyframe generation. | undefined |
| Resolusi | Enum (opsi 3) | - | 768p |
| Durasi (detik) | Enum (opsi 11) | durasi keluaran () . | 5 |
| Prompt Expansion Mode | Enum (opsi 3) | Prompt rewriting effort. 'quality' can add up to ~30s of processing. | balanced |
, .
The camera slowly pushes in as the subject turns toward the light and fabric moves in a soft breeze.Starting frame image. The output canvas follows this image.
undefinedOptional final frame image. Can be used alone for end-keyframe generation.
undefined-
768pdurasi keluaran () .
5Prompt rewriting effort. 'quality' can add up to ~30s of processing.
balancedDokumentasi pengembang
Provide a source frame: Set image_url to your starting image, or end_image_url alone to generate up to that frame. At least one of the two is required.
Describe the motion: Write prompt to describe the motion, action, and camera movement you want applied to the image.
Pick resolution and duration: Set resolution to 480p, 768p, or 1080p, and duration to any value from 5 to 15 seconds.
Choose prompt expansion (optional): Set prompt_expansion_mode to disabled, balanced (default), or quality for stronger prompt rewriting before generation.
Submit and retrieve: POST the request, then poll /predictions/{request_id}/result or supply a webhook_url to receive the finished video URL.
curl -X POST https://api.muapi.ai/api/v1/minimax-h3-max-image-to-video \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "The camera slowly pushes in as the subject turns toward the light and fabric moves in a soft breeze.",
"image_url": "https://example.com/portrait.jpg",
"resolution": "1080p",
"duration": 6
}'
Sering ditanyakan
It animates a source image into a video clip, using a text prompt to describe the desired motion and camera movement, at up to 1080p resolution.
No. You need at least one of `image_url` or `end_image_url` — using only `end_image_url` generates motion leading up to that frame.
Resolution can be 480p, 768p, or 1080p, and duration can be any value from 5 to 15 seconds.
It controls how much the model rewrites your motion prompt before rendering. `disabled` uses your prompt as-is, `balanced` applies light rewriting, and `quality` applies the most thorough rewrite.
Poll the request ID returned by the initial POST at `/predictions/{request_id}/result`, or pass a `webhook_url` to be notified when the video is ready.
Yes, supply `end_image_url` alongside or instead of `image_url` to anchor how the clip ends.