Use the MiniMax H3 Max API on Muapi for high-fidelity text-to-video, image-to-video, and reference-to-video generation at up to 1080p, plus a faster Turbo mode for rapid iteration. Each workflow uses Muapi's unified asynchronous API pattern, so your application can submit a request, track its status, and retrieve the completed video with one API key.
MiniMax H3 Max is a higher-fidelity Hailuo video model with output up to 1080p, prompt expansion, and text-to-video, image-to-video, and reference-to-video workflows. A Turbo mode trades some quality for a lower per-second rate. Use Muapi's standard asynchronous request and polling flow to manage every H3 Max generation with one API key.
Generate high-fidelity video up to 1080p directly from a detailed text prompt covering scene composition, action, camera direction, and style.
Animate a supplied start and/or end frame with a motion prompt at up to 1080p, keeping the source composition as the visual anchor.
Generate video from a prompt plus up to 12 combined image and video references and optional audio, for stronger subject and style consistency at up to 1080p.
A faster, lower-cost text-to-video mode of H3 Max for rapid prompt iteration and high-volume generation, still up to 1080p.
A faster, lower-cost image-to-video mode of H3 Max for rapid motion-prompt iteration and high-volume animation, still up to 1080p.
MiniMax H3 Max is a higher-fidelity evolution of the [MiniMax H3](/minimax-h3) video family, available through Muapi's unified API. It renders at up to 1080p, supports an optional prompt-expansion step that rewrites your prompt for stronger cinematic detail, and covers three workflows: start from a written idea, animate a supplied image, or guide a new video with multimodal references. A Turbo mode shares the same request shape at roughly half the per-second cost, for teams that need to iterate on prompts quickly before committing to a final render.
For teams building creative products, marketing tools, video editors, or automated media workflows, H3 Max provides a practical split between open-ended generation, tightly guided shots, and fast/cheap iteration. Text to Video is the right starting point when the scene exists only in a brief. Image to Video turns approved key art or product photography into motion, with optional start and end frame control. Reference to Video combines a prompt with images, clips, and optional audio to preserve creative direction you already have.
Muapi standardizes authentication, asynchronous task responses, webhooks, and result retrieval across the text-to-video and image-to-video catalog. MiniMax H3 Max can therefore sit alongside other video models, including the standard-resolution [MiniMax H3](/minimax-h3) family, in a routing or evaluation workflow without a new provider account or a second response format.
Use a descriptive prompt to define the subject, setting, action, camera direction, lighting, pacing, and visual style at up to 1080p. This is the flexible mode for concepts, storyboards, and new scenes.
Start with an image that already contains the subject, art direction, and composition you want, with optional start and end frame control. Focus the prompt on motion and camera movement.
Supply up to 12 combined reference images and videos, plus optional audio, to guide subject consistency, motion, visual style, and scene continuity.
Render the same text-to-video request shape at roughly half the per-second cost of standard H3 Max, ideal for rapid prompt iteration before a final render.
Animate a source image at roughly half the per-second cost of standard H3 Max, ideal for rapid motion-prompt iteration before a final render.
Create high-fidelity product-motion concepts, social ads, and visual variations from campaign briefs or approved product photography.
Turn scripts and storyboards into fast animatics, camera references, and transition studies before a live shoot.
Generate vertical-ready motion concepts, visual hooks, and short scene cutaways for creator workflows, iterating quickly with Turbo before a final H3 Max render.
Add high-fidelity video generation to a design, presentation, ecommerce, or agent-driven application through one API contract.
The strongest video prompts establish a clear subject, place, movement, camera choice, and atmosphere. Give motion a reason: a gust of wind moves fabric, a tracking camera follows a runner, or a crane shot reveals the scale of a landscape.
For Image to Video, retain the visual facts already in the image and describe only the intended movement. For Reference to Video, name how each reference should influence the subject, style, motion, or soundtrack. Set prompt_expansion_mode to quality when a prompt needs more cinematic detail than you wrote yourself.
A macro tracking shot follows a glass perfume bottle on blue silk as morning light moves across the label. The camera makes a slow 180-degree orbit; translucent fabric ripples in a gentle breeze, with soft editorial studio lighting and shallow depth of field.
Every MiniMax H3 Max request is asynchronous. Submit the generation request and store the returned request ID. Your application can poll for the final media URL or provide a webhook URL when submitting the task. This keeps long-running video work off the request thread and matches the rest of Muapi's generative-media APIs.
Submit a prompt to an H3 Max endpoint, receive a request ID, and poll the standard prediction result endpoint for the completed video.
curl -X POST https://api.muapi.ai/api/v1/minimax-h3-max-text-to-video \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"A cinematic tracking shot of a silver sports car driving through a rain-soaked city at night","aspect_ratio":"16:9","resolution":"1080p","duration":6}'
# Response: {"request_id":"REQUEST_ID"}Start with minimax-h3-max-text-to-video, or use the image, reference, and turbo endpoints for guided or faster workflows.
curl https://api.muapi.ai/api/v1/predictions/REQUEST_ID/result \ -H "x-api-key: YOUR_API_KEY" # The completed response contains the generated video URL.
The response contains the task status and generated output URL when the video is complete.
Create high-fidelity videos up to 1080p from a text prompt, animate a source image with optional start/end frame control, or guide a new video with image, video, and optional audio references.
H3 Max renders at up to 1080p and supports an optional prompt-expansion step. It also adds Turbo text-to-video and image-to-video modes at roughly half the per-second cost for fast iteration.
Use Text to Video for an original scene, Image to Video when you need to animate a supplied composition, Reference to Video when multiple assets should guide the output, and Turbo for cheap, fast iteration before a final render.
Store the request ID returned by the submit call. Poll the prediction result endpoint from a worker or receive completion through a webhook instead of keeping a browser request open.
480p, 768p, and 1080p resolutions, with durations from 5 to 15 seconds across every H3 Max workflow.
MiniMax H3 Max uses Muapi's pay-per-generation billing model, with cost scaling by resolution and duration. Review the model page for current pricing on the selected workflow.
No. MiniMax H3 Max uses your existing Muapi API key, together with the same task polling and webhook pattern as other Muapi models.
Use one Muapi API key for high-fidelity text-to-video, image-to-video, reference-to-video, and fast Turbo workflows.