# MiniMax H3 Max Turbo > Generate 5-15s AI videos from text or a first-frame image with MiniMax H3 Max Turbo, with strong prompt adherence. ## Overview - **Endpoint**: `https://api.segmind.com/v1/minimax-h3-max-turbo` - **Model ID**: `minimax-h3-max-turbo` - **Category**: Text-to-Video Generation - **Type**: Synchronous (Direct response) - **Average Latency**: ~18.7s (30-day average) - **Average Cost**: $0.3316 per run (observed across past runs, not a price — see Pricing) - **Provider**: FAL ## Pricing | resolution | rate | | --- | --- | | 480P | $0.025 per second | | 768P | $0.04 per second | Billed per second of output video at the selected resolution. Same rate for text-to-video and image-to-video. ## API Information This model uses a **synchronous response pattern**: 1. Make a POST request with your parameters 2. Receive the output directly in the response (binary for images/videos/audio, JSON for text) 3. No polling required - response is immediate ### Input Schema The API accepts the following input parameters: - **`prompt`** (`string`, _required_): Prompt Describe the scene, subject, motion and camera. Used for both text-to-video and image-to-video. - **`image`** (`File (URL)`, _optional_): Image (optional first frame) Optional first frame. When set, the video starts from this image and takes its aspect ratio (image-to-video). Leave empty for text-to-video. - **`last_frame_image`** (`File (URL)`, _optional_): Last Frame Image Optional end frame, used together with Image: the model interpolates from the first frame to this one. - **`duration`** (`integer`, _optional_): Duration (seconds) Length of the generated video, 5 to 15 seconds. Billed per second of output. - Default: `5` - Range: 5 to 15 - **`resolution`** (`string`, _optional_): Resolution Output resolution. 480P is the cheaper tier; 768P is the default quality tier. - Default: `"768P"` - Options: "480P" (480P), "768P" (768P) - **`aspect_ratio`** (`string`, _optional_): Aspect Ratio Output aspect ratio for text-to-video. Ignored when an Image is provided (the output follows the image). - Default: `"16:9"` - Options: "21:9" (21:9), "16:9" (16:9), "4:3" (4:3), "1:1" (1:1), "3:4" (3:4), "9:16" (9:16) - **`prompt_expansion_mode`** (`string`, _optional_): Prompt Expansion How much effort to spend rewriting the prompt before generation. 'balanced' returns in about a second; 'quality' spends up to ~30s on a richer prompt. - Default: `"balanced"` - Options: "balanced" (Balanced), "quality" (Quality) - **`seed`** (`integer`, _optional_): Seed Seed for reproducible generation. -1 for random. - Default: `-1` - Range: -1 to 999999999999999 **Required Parameters Example**: ```json { "prompt": "A track cyclist in a red skinsuit sprints off the final banking of a wooden velodrome, rising out of the saddle as the bike surges forward; the camera tracks level with the front wheel while the pine boards blur past behind." } ``` **Full Example**: ```json { "prompt": "A track cyclist in a red skinsuit sprints off the final banking of a wooden velodrome, rising out of the saddle as the bike surges forward; the camera tracks level with the front wheel while the pine boards blur past behind.", "image": "https://example.com/image.jpg", "last_frame_image": "https://example.com/image.jpg", "duration": 6, "resolution": "768P", "aspect_ratio": "16:9", "prompt_expansion_mode": "balanced", "seed": -1 } ``` ### Output Schema The API returns a synchronous response based on the model type: **For Image/Video/Audio Models**: - Response contains binary data (image/png, video/mp4, audio/mp3) - Content-Type header indicates the media type - Save the response body directly to a file **For Text Models**: - Response is JSON with the generated text - Structure varies by model **HTTP Response Codes**: - **200 - OK**: Request successful, output in response body - **400 - Bad Request**: Invalid parameters - **401 - Unauthorized**: Invalid or missing API key - **404 - Not Found**: Model not found - **406 - Not Acceptable**: Insufficient credits - **429 - Too Many Requests**: Rate limit exceeded - **500 - Server Error**: Internal server error ## About ### MiniMax H3 Max Turbo: Text-to-Video and Image-to-Video #### What is MiniMax H3 Max Turbo? MiniMax H3 Max Turbo is a fast AI video generator that turns a text prompt or a first-frame image into a 5 to 15 second clip at 480P or 768P. It is the speed-tuned tier of H3 Max — fal's post-trained version of MiniMax's H3 (Hailuo 3) video model — and renders faster than real time while holding H3 Max-level quality. If you know MiniMax H3 or Hailuo, Turbo is the tier for exploring many shots quickly rather than mastering at 2K. #### Key Features - **Strong prompt adherence and aesthetics.** Follows specific, multi-clause briefs — subject, action and camera move — closely and cleanly. - **Text-to-video and image-to-video in one endpoint.** Set the image parameter to animate a still first frame. - **First-and-last-frame keyframing.** Supply a first and last frame and the model interpolates the motion between them. - **Legible on-screen text and kinetic typography.** Renders correctly spelled titles, captions and brand marks. - **Flexible output.** 5 to 15 seconds, 480P or 768P, six aspect ratios from 21:9 to 9:16. - **Speed.** Roughly twice as fast as H3 Max. #### Best Use Cases Turbo suits rapid iteration: ad concepts, social short-form, product motion and motion-graphics titles. In testing it realized a detailed sports brief (a track cyclist coming off a velodrome banking with the camera on the front wheel), animated a product still into a chocolate pour, interpolated a first-and-last-frame keyframe pair, and rendered a legible motion title with native synchronized audio in the returned mp4. #### Prompt Tips and Output Quality Write one clear subject, one action and one motivated camera move; the model rewards concrete nouns over adjectives. For image-to-video, feed a first frame that already contains something poised to move. Keep prompt_expansion_mode on balanced to honor your exact wording, or switch to quality to enrich short prompts. Output was photoreal and temporally stable at 768P. ## Usage Guide ### How to Use MiniMax H3 Max Turbo MiniMax H3 Max Turbo generates 5 to 15 second videos from a text prompt or a first-frame image. It rewards precise briefs: name the subject, the single action, and one camera move. Use text-to-video for original shots, image-to-video to animate a still, and first-and-last-frame keyframing for controlled transitions. #### Recommended Settings - **Text-to-video:** 768P, duration 5 to 6s (up to 15s), aspect_ratio 16:9 or 9:16, prompt_expansion_mode balanced, seed -1. - **Image-to-video:** set image to your first frame; aspect_ratio is ignored (output follows the image). Start at 5s. - **First-and-last-frame:** set image and last_frame_image to two frames of the same subject and background; duration 5s. - **Drafts vs finals:** 480P for exploration, 768P for delivery. Use quality expansion only to enrich short prompts. #### Example prompts to try Each example below ran end to end at the settings above; the length, motion and framing carried straight from the prompt into the finished clip. - **Prompt-adhering fast motion (default).** Prompt: "A track cyclist in a red skinsuit sprints off the final banking of a wooden velodrome, rising out of the saddle as the bike surges forward; the camera tracks level with the front wheel while the pine boards blur past behind." The clip realized every clause with clean, fast, legible motion. Tip: write one subject plus one action plus one motivated camera move; 768P at 6s. - **Image-to-video from a first frame.** Prompt: "A thin ribbon of dark chocolate pours from the copper jug and spreads glossily across the top of the tart, its surface rippling and settling as the jug tilts back and lifts away." The supplied chocolate-tart still came to life. Tip: feed a first frame that already holds something poised to move; i2v ignores aspect_ratio and follows the image. - **Legible kinetic typography.** Prompt: "Bold cream sans-serif letters spelling 'NORTH LOOP 7' snap onto a flat charcoal background one word at a time, a magenta bar wipes left to right beneath the words, then the whole title compresses slightly and settles sharp and centered." Cream text snapped in word by word, correctly spelled, with a magenta underline. Tip: put the exact words in quotes and name the beats; text stays crisp at 768P. ##### Reference-guided: first-and-last-frame This mode takes two frames — set image as the first frame and last_frame_image as the end frame, keeping the same subject and background so only the pose changes. - **First-and-last-frame keyframing.** Prompt: "The folded paper crane rises and straightens on the walnut table, its creased wings spreading outward and lifting until it stands upright, sharp folds catching the light." The crane's wings visibly rose and spread as it straightened between the two frames. Tip: change only the pose you want animated. Also built for prompts like "a lathe peels a bright curl of steel from a spinning bar" or "a neon sign flickers on above a diner counter." ## FAQ ### What is the maximum resolution? 768P — for 2K, use base MiniMax H3. ### Does it support image-to-video? Yes. Set the image parameter for a first frame, and optionally last_frame_image for a first-and-last-frame transition. ### How long can a clip be? From 5 to 15 seconds, at 24 fps. ### Does it generate audio? Base H3 produces native synchronized audio, and the returned mp4 carried an audio track in testing. ### How is it different from MiniMax H3? It is the fast Turbo tier of fal's H3 Max, built on MiniMax H3, tuned for speed and prompt adherence rather than 2K mastering. ## Example Inputs Parameter sets that produced real outputs on the playground. ### Example 1 ```json { "prompt": "A track cyclist in a red skinsuit sprints off the final banking of a wooden velodrome, rising out of the saddle as the bike surges forward; the camera tracks level with the front wheel while the pine boards blur past behind.", "duration": 6, "resolution": "768P", "aspect_ratio": "16:9", "prompt_expansion_mode": "balanced", "seed": -1 } ``` Output: https://segmind-resources.s3.amazonaws.com/output/minimax-h3-max-turbo-example-output-9b000392.mp4 ### Example 2 ```json { "prompt": "A thin ribbon of dark chocolate pours from the copper jug and spreads glossily across the top of the tart, its surface rippling and settling as the jug tilts back and lifts away.", "image": "https://segmind-resources.s3.amazonaws.com/input/minimax-h3-max-turbo-cardA-choc-firstframe-97907ef7.jpg", "duration": 5, "resolution": "768P", "aspect_ratio": "16:9", "prompt_expansion_mode": "balanced", "seed": -1 } ``` Output: https://segmind-resources.s3.amazonaws.com/output/minimax-h3-max-turbo-example-image-to-video-choc-pour-20fecc6f.mp4 ### Example 3 ```json { "prompt": "Bold cream sans-serif letters spelling 'NORTH LOOP 7' snap onto a flat charcoal background one word at a time, a magenta bar wipes left to right beneath the words, then the whole title compresses slightly and settles sharp and centered.", "duration": 5, "resolution": "768P", "aspect_ratio": "16:9", "prompt_expansion_mode": "balanced", "seed": -1 } ``` Output: https://segmind-resources.s3.amazonaws.com/output/minimax-h3-max-turbo-example-kinetic-typography-183d4093.mp4 ## Usage Examples ### cURL ```bash curl -X POST "https://api.segmind.com/v1/minimax-h3-max-turbo" \ -H "x-api-key: YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "prompt": "A track cyclist in a red skinsuit sprints off the final banking of a wooden velodrome, rising out of the saddle as the bike surges forward; the camera tracks level with the front wheel while the pine boards blur past behind.", "image": "https://example.com/image.jpg", "last_frame_image": "https://example.com/image.jpg", "duration": 6, "resolution": "768P", "aspect_ratio": "16:9", "prompt_expansion_mode": "balanced", "seed": -1 }' ``` ### Python ```python import requests import json api_key = "YOUR_API_KEY" url = "https://api.segmind.com/v1/minimax-h3-max-turbo" data = { "prompt": "A track cyclist in a red skinsuit sprints off the final banking of a wooden velodrome, rising out of the saddle as the bike surges forward; the camera tracks level with the front wheel while the pine boards blur past behind.", "image": "https://example.com/image.jpg", "last_frame_image": "https://example.com/image.jpg", "duration": 6, "resolution": "768P", "aspect_ratio": "16:9", "prompt_expansion_mode": "balanced", "seed": -1 } response = requests.post( url, json=data, headers={ 'x-api-key': api_key, 'Content-Type': 'application/json' } ) if response.status_code == 200: # For image/video/audio models, response.content contains the binary data with open('output.png', 'wb') as f: f.write(response.content) print('Generation complete, saved to output.png') else: print(f"Error: {response.status_code}") print(response.text) ``` ### JavaScript ```javascript const apiKey = 'YOUR_API_KEY'; const url = 'https://api.segmind.com/v1/minimax-h3-max-turbo'; const data = { "prompt": "A track cyclist in a red skinsuit sprints off the final banking of a wooden velodrome, rising out of the saddle as the bike surges forward; the camera tracks level with the front wheel while the pine boards blur past behind.", "image": "https://example.com/image.jpg", "last_frame_image": "https://example.com/image.jpg", "duration": 6, "resolution": "768P", "aspect_ratio": "16:9", "prompt_expansion_mode": "balanced", "seed": -1 }; const response = await fetch(url, { method: 'POST', headers: { 'x-api-key': apiKey, 'Content-Type': 'application/json', }, body: JSON.stringify(data), }); if (response.ok) { // For image/video/audio models, response contains binary data const blob = await response.blob(); const downloadUrl = URL.createObjectURL(blob); // Create download link const a = document.createElement('a'); a.href = downloadUrl; a.download = 'output.png'; a.click(); console.log('Generation complete'); } ``` ## Additional Resources ### Documentation - [Model Playground](https://www.segmind.com/models/minimax-h3-max-turbo) - [API Documentation](https://www.segmind.com/models/minimax-h3-max-turbo/api) - [Pricing Details](https://www.segmind.com/models/minimax-h3-max-turbo/pricing) - [Platform Documentation](https://docs.segmind.com/)