Call MiniMax H3 (open weights hosted by WaveSpeedAI) and the official MiniMax H3, H3 Lite, H3 Highspeed and H3 Image endpoints from Python with one function per endpoint: text-to-video, image-to-video, reference-to-video, video editing, video extension, motion control, text-to-image, image editing. Built on the official wavespeed SDK and the WaveSpeedAI hosted API — no GPU, pay per generation.
- MiniMax H3 API guide, live pricing and playground: wavespeed.ai/minimax-h3-api
- No code: generate in the browser with the AI Video Generator on WaveSpeedAI
pip install git+https://github.com/WaveSpeedAI/minimax-h3-api.git
export WAVESPEED_API_KEY="your-api-key" # https://wavespeed.ai/accesskeyimport minimax_h3_api
urls = minimax_h3_api.text_to_video(
prompt="A slow cinematic dolly shot through a rain-soaked neon night market, reflections on the pavement",
)
print(urls[0])Every function waits for the generation to finish and returns a list of output URLs. Optional arguments you leave out are not sent, so the API defaults apply. Inputs such as images, videos and audio are passed as URLs; use minimax_h3_api.upload("local-file") to get a URL for a local file.
Each model page has a playground, live pricing and copy-paste API snippets. Prices are per generation and depend on the settings you choose; check a request before running it with wavespeed price <model> from the CLI.
wavespeed-ai/minimax-h3/text-to-video — Open-weights MiniMax H3 text-to-video on WaveSpeed infrastructure: coherent 480p/540p/768p/1080p videos with native stereo audio from a single prompt, 5-15 second durations, flexible aspect ratios, and affordable per-second pricing.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the video scene, action, camera movement, and the desired soundtrack. Audio is generated natively together with the video. | ||
aspect_ratio |
str | "16:9" |
16:9, 9:16, 1:1, 4:3, 3:4, 21:9, 9:21 |
Output aspect ratio. | |
resolution |
str | "768p" |
480p, 540p, 768p, 1080p |
Output video resolution. 768p is the model's native canvas; 480p is a faster, lower-cost tier; 540p is a mid tier at 1.5x the 480p price. 1080p is the highest-quality full-HD tier at 2x the 768p price (generation takes longer). | |
duration |
int | 5 |
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. | |
enable_prompt_expansion |
bool | true |
Expand the prompt into a detailed description before generation when that helps. Turn off to use the prompt exactly as written. | ||
seed |
int | The random seed to use for the generation. -1 means a random seed will be used. |
wavespeed-ai/minimax-h3/text-to-video-lora — Open-weights MiniMax H3 text-to-video on WaveSpeed infrastructure: coherent 480p/540p/768p/1080p videos with native stereo audio from a single prompt, 5-15 second durations, flexible aspect ratios, and affordable per-second pricing, with custom LoRA support.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the video scene, action, camera movement, and the desired soundtrack. Audio is generated natively together with the video. | ||
aspect_ratio |
str | "16:9" |
16:9, 9:16, 1:1, 4:3, 3:4, 21:9, 9:21 |
Output aspect ratio. | |
resolution |
str | "768p" |
480p, 540p, 768p, 1080p |
Output video resolution. 768p is the model's native canvas; 480p is a faster, lower-cost tier; 540p is a mid tier at 1.5x the 480p price. 1080p is the highest-quality full-HD tier at 2x the 768p price (generation takes longer). | |
duration |
int | 5 |
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. | |
loras |
List[Dict[str, Any]] | up to 3 items | List of LoRAs to apply (max 3). | ||
seed |
int | The random seed to use for the generation. -1 means a random seed will be used. |
wavespeed-ai/minimax-h3/image-to-video — Open-weights MiniMax H3 image-to-video on WaveSpeed infrastructure: animate a first-frame image (optionally with a last frame) into a coherent 480p/540p/768p/1080p video with native stereo audio, 5-15 second durations, and affordable per-second pricing.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the desired motion, scene, and soundtrack. Audio is generated natively together with the video. | ||
image |
str | yes | First-frame image URL. The output canvas follows this image's aspect ratio. | ||
last_image |
str | Optional last-frame image URL. When provided, the video interpolates from the first frame to this frame. | |||
resolution |
str | "768p" |
480p, 540p, 768p, 1080p |
Output video resolution. 768p is the model's native canvas; 480p is a faster, lower-cost tier; 540p is a mid tier at 1.5x the 480p price. 1080p is the highest-quality full-HD tier at 2x the 768p price (generation takes longer). | |
duration |
int | 5 |
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. | |
enable_prompt_expansion |
bool | true |
Expand the prompt into a detailed description before generation when that helps. Turn off to use the prompt exactly as written. | ||
seed |
int | The random seed to use for the generation. -1 means a random seed will be used. |
wavespeed-ai/minimax-h3/image-to-video-lora — Open-weights MiniMax H3 image-to-video on WaveSpeed infrastructure: animate a first-frame image (optionally with a last frame) into a coherent 480p/540p/768p/1080p video with native stereo audio, 5-15 second durations, and affordable per-second pricing, with custom LoRA support.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the desired motion, scene, and soundtrack. Audio is generated natively together with the video. | ||
image |
str | yes | First-frame image URL. The output canvas follows this image's aspect ratio. | ||
last_image |
str | Optional last-frame image URL. When provided, the video interpolates from the first frame to this frame. | |||
resolution |
str | "768p" |
480p, 540p, 768p, 1080p |
Output video resolution. 768p is the model's native canvas; 480p is a faster, lower-cost tier; 540p is a mid tier at 1.5x the 480p price. 1080p is the highest-quality full-HD tier at 2x the 768p price (generation takes longer). | |
duration |
int | 5 |
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. | |
loras |
List[Dict[str, Any]] | up to 3 items | List of LoRAs to apply (max 3). | ||
seed |
int | The random seed to use for the generation. -1 means a random seed will be used. |
wavespeed-ai/minimax-h3/reference-to-video — Open-weights MiniMax H3 reference-to-video on WaveSpeed infrastructure: generate coherent 480p/540p/768p/1080p videos with native stereo audio guided by up to 9 reference images, 3 reference videos, and 3 reference audios.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the desired video. Refer to reference inputs as <Picture 1>..<Picture 9>, <Video 1>..<Video 3>, and <Audio 1>..<Audio 3>. Audio is generated natively together with the video. | ||
reference_images |
List[str] | up to 9 items | Reference image URLs (up to 9). At least one reference input is required. | ||
reference_videos |
List[str] | up to 3 items | Reference video URLs (up to 3, usable at any output resolution). Their soundtracks are used automatically. Total reference video duration is budgeted to 15 seconds; longer inputs are truncated fairly. | ||
reference_audios |
List[str] | up to 3 items | Standalone reference audio URLs (up to 3, trimmed to 15 seconds each). | ||
aspect_ratio |
str | "16:9" |
16:9, 9:16, 1:1, 4:3, 3:4, 21:9, 9:21 |
Output aspect ratio. | |
resolution |
str | "768p" |
480p, 540p, 768p, 1080p |
Output video resolution. 768p is the model's native canvas; 480p is a faster, lower-cost tier; 540p is a mid tier at 1.5x the 480p price. 1080p is the highest-quality full-HD tier at 2x the 768p price (generation takes longer). | |
duration |
int | 5 |
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. | |
enable_prompt_expansion |
bool | true |
Expand the prompt into a detailed description before generation when that helps. Turn off to use the prompt exactly as written. | ||
seed |
int | The random seed to use for the generation. -1 means a random seed will be used. |
wavespeed-ai/minimax-h3/reference-to-video-lora — Open-weights MiniMax H3 reference-to-video on WaveSpeed infrastructure: generate coherent 480p/540p/768p/1080p videos with native stereo audio guided by up to 9 reference images, 3 reference videos, and 3 reference audios, with custom LoRA support.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the desired video. Refer to reference inputs as <Picture 1>..<Picture 9>, <Video 1>..<Video 3>, and <Audio 1>..<Audio 3>. Audio is generated natively together with the video. | ||
reference_images |
List[str] | up to 9 items | Reference image URLs (up to 9). At least one reference input is required. | ||
reference_videos |
List[str] | up to 3 items | Reference video URLs (up to 3, usable at any output resolution). Their soundtracks are used automatically. Total reference video duration is budgeted to 15 seconds; longer inputs are truncated fairly. | ||
reference_audios |
List[str] | up to 3 items | Standalone reference audio URLs (up to 3, trimmed to 15 seconds each). | ||
aspect_ratio |
str | "16:9" |
16:9, 9:16, 1:1, 4:3, 3:4, 21:9, 9:21 |
Output aspect ratio. | |
resolution |
str | "768p" |
480p, 540p, 768p, 1080p |
Output video resolution. 768p is the model's native canvas; 480p is a faster, lower-cost tier; 540p is a mid tier at 1.5x the 480p price. 1080p is the highest-quality full-HD tier at 2x the 768p price (generation takes longer). | |
duration |
int | 5 |
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. | |
loras |
List[Dict[str, Any]] | up to 3 items | List of LoRAs to apply (max 3). | ||
seed |
int | The random seed to use for the generation. -1 means a random seed will be used. |
wavespeed-ai/minimax-h3/video-edit — MiniMax H3 Open Weights Video-Edit edits an input video from a natural-language prompt. The input video drives subject identity, composition, and motion while the model rewrites lighting, style, weather, environment, or specific elements as instructed, with native stereo audio generated in the same pass.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Describe the edit you want applied to the input video - change lighting, weather, style, environment, or specific elements. | ||
video |
str | yes | URL of the input video to edit. It drives subject identity, composition, and motion while the prompt rewrites lighting, style, environment, or specific elements. | ||
reference_images |
List[str] | up to 9 items | Optional reference image URLs to guide the edit (subject identity, style, etc.). Refer to them in the prompt as <Picture 1>..<Picture 9>. | ||
reference_audios |
List[str] | up to 3 items | Optional reference audio URLs to guide audio generation. Refer to them in the prompt as <Audio 1>..<Audio 3>. | ||
resolution |
str | "768p" |
480p, 540p, 768p, 1080p |
Output video resolution. 768p is the model's native canvas; 480p is a faster, lower-cost tier; 540p is a mid tier at 1.5x the 480p price. 1080p is the highest-quality full-HD tier at 2x the 768p price (generation takes longer). | |
aspect_ratio |
str | 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, 9:21 |
Aspect ratio of the output video. Adapts to the input video if not specified. | ||
duration |
int | 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. Auto-detected from the input video if not specified. | ||
generate_audio |
bool | true |
Whether to generate native audio for the edited output. When set to false, the input video's audio track is preserved on the output instead. | ||
enable_prompt_expansion |
bool | true |
Expand the prompt into a detailed description before generation when that helps. Turn off to use the prompt exactly as written. | ||
seed |
int | The random seed to use for the generation. A negative value means a random seed will be used. |
wavespeed-ai/minimax-h3/video-extend — MiniMax H3 Open Weights Video-Extend appends a new cinematic continuation to an existing video. A fresh segment with native stereo audio is generated from the input video's last frame and a natural-language prompt, then concatenated onto the original.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the continuation - motion, scene, and soundtrack. Audio is generated natively together with the video. | ||
video |
str | yes | URL of the video to extend. The new segment is generated from its last frame and appended to the original. | ||
last_image |
str | Optional target last-frame image URL. When provided, the new segment interpolates from the input video's last frame to this image. | |||
resolution |
str | "768p" |
480p, 540p, 768p, 1080p |
Resolution of the new segment. 768p is the model's native canvas; 480p is a faster, lower-cost tier; 540p is a mid tier at 1.5x the 480p price. 1080p is the highest-quality full-HD tier at 2x the 768p price (generation takes longer). | |
duration |
int | 5 |
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Length in seconds of the new segment to append. | |
enable_prompt_expansion |
bool | true |
Expand the prompt into a detailed description before generation when that helps. Turn off to use the prompt exactly as written. | ||
seed |
int | The random seed to use for the generation. A negative value means a random seed will be used. |
wavespeed-ai/minimax-h3/controlnet-union — MiniMax H3 Open Weights ControlNet Union generates a new video that follows the motion and composition of a source video. Pose, depth, edges, lines, scribble or grayscale structure is extracted from the source automatically and guides the output, optionally with reference images for the subject or style, with native stereo audio generated in the same pass.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Describe the output video. The source video's motion and composition are carried over through the selected control. Refer to reference inputs as <Picture 1>..<Picture 9> and <Audio 1>..<Audio 3>. | ||
video |
str | yes | URL of the source video (up to 15 seconds are used). The control signal is extracted from it automatically and drives motion and composition; the output keeps its duration and aspect ratio. | ||
control_type |
str | "pose" |
pose, depth, canny, soft_edge, lines, scribble, gray |
What to keep from the source video: pose (body, hand and face motion), depth (spatial layout), canny (precise edges), soft_edge (softer outlines), lines (straight lines, for architecture and interiors), scribble (loose sketch lines), gray (brightness structure, for recoloring). | |
control_strength |
float | 1 |
0–2 | How strongly the control signal constrains the output. Lower values give the prompt more freedom. | |
reference_images |
List[str] | up to 9 items | Optional reference image URLs, e.g. the character to animate or a target style. Refer to them in the prompt as <Picture 1>..<Picture 9>. | ||
reference_audios |
List[str] | up to 3 items | Optional reference audio URLs to guide audio generation. Refer to them in the prompt as <Audio 1>..<Audio 3>. | ||
resolution |
str | "768p" |
480p, 540p, 768p, 1080p |
Output video resolution. 768p is the model's native canvas; 480p is a faster, lower-cost tier; 540p is a mid tier at 1.5x the 480p price. 1080p is the highest-quality full-HD tier at 2x the 768p price (generation takes longer). | |
generate_audio |
bool | true |
Whether to generate native audio. When set to false, the source video's audio track is preserved on the output instead. | ||
seed |
int | The random seed to use for the generation. A negative value means a random seed will be used. |
wavespeed-ai/minimax-h3/text-to-image — MiniMax H3 Open Weights Text-to-Image generates photorealistic, cinema-grade stills from a text prompt at 1K or 2K resolution with a wide choice of aspect ratios.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | The positive prompt for the generation. | ||
aspect_ratio |
str | "1:1" |
1:1, 1:2, 2:1, 1:3, 3:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 9:21, 21:9 |
The aspect ratio of the generated image. Defaults to 1:1. | |
resolution |
str | "1k" |
1k, 1.5k, 2k |
Total output resolution tier: 1k (~1 megapixel), 1.5k (~2.4 megapixels) or 2k (~4 megapixels). | |
output_format |
str | "jpeg" |
jpeg, png, webp |
The format of the output image. | |
enable_prompt_expansion |
bool | true |
Expand the prompt into a detailed description before generation when that helps. Turn off to use the prompt exactly as written. | ||
seed |
int | The random seed to use for the generation. -1 means a random seed will be used. |
wavespeed-ai/minimax-h3/text-to-image-lora — MiniMax H3 Open Weights Text-to-Image generates photorealistic, cinema-grade stills from a text prompt at 1K or 2K resolution with a wide choice of aspect ratios. Loads up to 3 custom LoRA weights per request.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | The positive prompt for the generation. | ||
aspect_ratio |
str | "1:1" |
1:1, 1:2, 2:1, 1:3, 3:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 9:21, 21:9 |
The aspect ratio of the generated image. Defaults to 1:1. | |
resolution |
str | "1k" |
1k, 1.5k, 2k |
Total output resolution tier: 1k (~1 megapixel), 1.5k (~2.4 megapixels) or 2k (~4 megapixels). | |
output_format |
str | "jpeg" |
jpeg, png, webp |
The format of the output image. | |
loras |
List[Dict[str, Any]] | up to 3 items | List of LoRAs to apply (max 3). | ||
enable_prompt_expansion |
bool | true |
Expand the prompt into a detailed description before generation when that helps. Turn off to use the prompt exactly as written. | ||
seed |
int | The random seed to use for the generation. -1 means a random seed will be used. |
wavespeed-ai/minimax-h3/image-edit — MiniMax H3 Open Weights Image Edit re-renders a subject from up to 9 reference images into a new scene, outfit or style from a text instruction, preserving identity at 1K or 2K resolution.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | The edit instruction. Refer to the reference images as <Picture 1>, <Picture 2>, ... in order. Describe the new scene, outfit or style you want; the subject's identity is taken from the references. | ||
images |
List[str] | yes | 1–9 items | Reference images to edit from (1 to 9 URLs). The first image also sets the output aspect ratio unless aspect_ratio is given. | |
aspect_ratio |
str | 1:1, 1:2, 2:1, 1:3, 3:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 9:21, 21:9 |
The aspect ratio of the generated image. If not specified, it is determined automatically from the input image. | ||
resolution |
str | "1k" |
1k, 1.5k, 2k |
Total output resolution tier: 1k (~1 megapixel), 1.5k (~2.4 megapixels) or 2k (~4 megapixels). | |
output_format |
str | "jpeg" |
jpeg, png, webp |
The format of the output image. | |
seed |
int | The random seed to use for the generation. -1 means a random seed will be used. |
wavespeed-ai/minimax-h3/image-edit-lora — MiniMax H3 Open Weights Image Edit re-renders a subject from up to 9 reference images into a new scene, outfit or style from a text instruction, preserving identity at 1K or 2K resolution. Loads up to 3 custom LoRA weights per request.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | The edit instruction. Refer to the reference images as <Picture 1>, <Picture 2>, ... in order. Describe the new scene, outfit or style you want; the subject's identity is taken from the references. | ||
images |
List[str] | yes | 1–9 items | Reference images to edit from (1 to 9 URLs). The first image also sets the output aspect ratio unless aspect_ratio is given. | |
aspect_ratio |
str | 1:1, 1:2, 2:1, 1:3, 3:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 9:21, 21:9 |
The aspect ratio of the generated image. If not specified, it is determined automatically from the input image. | ||
resolution |
str | "1k" |
1k, 1.5k, 2k |
Total output resolution tier: 1k (~1 megapixel), 1.5k (~2.4 megapixels) or 2k (~4 megapixels). | |
output_format |
str | "jpeg" |
jpeg, png, webp |
The format of the output image. | |
loras |
List[Dict[str, Any]] | up to 3 items | List of LoRAs to apply (max 3). | ||
seed |
int | The random seed to use for the generation. -1 means a random seed will be used. |
minimax/h3/text-to-video — MiniMax H3 is a 2K text-to-video model that generates coherent videos from natural-language prompts with flexible duration and aspect-ratio controls.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the video scene, action, camera movement, and style. | ||
aspect_ratio |
str | "16:9" |
21:9, 16:9, 4:3, 1:1, 3:4, 9:16 |
Output aspect ratio. | |
resolution |
str | "768p" |
768p, 2k |
Output video resolution. | |
duration |
int | 5 |
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. |
minimax/h3/image-to-video — MiniMax H3 is a 2K image-to-video model that animates a first-frame image with natural-language motion and scene instructions, with optional last-frame control.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the desired motion and scene. | ||
image |
str | yes | First-frame image URL or Base64 data. Supported image dimensions are 256-5760 pixels per side, with an aspect ratio between 0.4 and 2.5. | ||
last_image |
str | Optional last-frame image URL or Base64 data. | |||
resolution |
str | "768p" |
768p, 2k |
Output video resolution. | |
duration |
int | 5 |
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. |
minimax/h3/reference-to-video — MiniMax H3 is a 2K reference-to-video model that combines reference images, videos, and audio with natural-language instructions to create coherent video scenes.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the desired video scene and interactions. | ||
reference_images |
List[str] | up to 9 items | Reference image URLs. At least one reference image or video is required. | ||
reference_videos |
List[str] | up to 3 items | Reference video URLs (up to 3). Each clip is normalized to 2-15 seconds and the combined duration is capped at 15 seconds. | ||
reference_audios |
List[str] | up to 3 items | Optional reference audio URLs. Audio cannot be provided alone. | ||
aspect_ratio |
str | "16:9" |
21:9, 16:9, 4:3, 1:1, 3:4, 9:16 |
Output aspect ratio. | |
resolution |
str | "768p" |
768p, 2k |
Output video resolution. | |
duration |
int | 5 |
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. |
minimax/h3-lite/text-to-video — MiniMax H3 Lite Text To Video generates 4-15 second videos at 768p or 480p.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the video scene, action, camera movement, and style. | ||
aspect_ratio |
str | "16:9" |
21:9, 16:9, 4:3, 1:1, 3:4, 9:16 |
Output aspect ratio. | |
resolution |
str | "768p" |
768p, 480p |
Output video resolution. | |
duration |
int | 5 |
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. |
minimax/h3-lite/image-to-video — MiniMax H3 Lite Image To Video generates 4-15 second videos at 768p or 480p.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the desired motion and scene. | ||
image |
str | yes | First-frame image URL or Base64 data. Supported image dimensions are 256-5760 pixels per side, with an aspect ratio between 0.4 and 2.5. | ||
last_image |
str | Optional last-frame image URL or Base64 data. | |||
resolution |
str | "768p" |
768p, 480p |
Output video resolution. | |
duration |
int | 5 |
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. |
minimax/h3-lite/reference-to-video — MiniMax H3 Lite Reference To Video generates 4-15 second videos at 768p or 480p.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the desired video scene and interactions. | ||
reference_images |
List[str] | up to 9 items | Up to 9 reference image URLs. At least one reference image or video is required. Images must be 256-5760 pixels per side, with a width-to-height ratio of 0.4-2.5. | ||
reference_videos |
List[str] | up to 3 items | Up to 3 reference video URLs. Clips are adjusted to 2-15 seconds each, with a combined maximum of 15 seconds; longer clips are trimmed and shorter clips are extended. Use MP4 or MOV, up to 50 MB per file, 256-5760 pixels per side, width-to-height ratio 0.4-2.5, and 23.976-60 FPS. At least one image or video is required. | ||
reference_audios |
List[str] | up to 3 items | Up to 3 optional WAV or MP3 reference audio URLs, up to 15 MB per file. Each clip must be 2-15 seconds, with a combined maximum of 15 seconds. Audio is not automatically trimmed and cannot be provided alone. | ||
aspect_ratio |
str | "16:9" |
21:9, 16:9, 4:3, 1:1, 3:4, 9:16 |
Output aspect ratio. | |
resolution |
str | "768p" |
768p, 480p |
Output video resolution. | |
duration |
int | 5 |
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. |
minimax/h3-highspeed/text-to-video — MiniMax H3 Highspeed Text To Video generates 4-15 second videos at 768p.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the video scene, action, camera movement, and style. | ||
aspect_ratio |
str | "16:9" |
21:9, 16:9, 4:3, 1:1, 3:4, 9:16 |
Output aspect ratio. | |
resolution |
str | "768p" |
768p |
Output video resolution. | |
duration |
int | 5 |
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. |
minimax/h3-highspeed/image-to-video — MiniMax H3 Highspeed Image To Video generates 4-15 second videos at 768p.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the desired motion and scene. | ||
image |
str | yes | First-frame image URL or Base64 data. Supported image dimensions are 256-5760 pixels per side, with an aspect ratio between 0.4 and 2.5. | ||
last_image |
str | Optional last-frame image URL or Base64 data. | |||
resolution |
str | "768p" |
768p |
Output video resolution. | |
duration |
int | 5 |
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. |
minimax/h3-highspeed/reference-to-video — MiniMax H3 Highspeed Reference To Video generates 4-15 second videos at 768p.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | Text description of the desired video scene and interactions. | ||
reference_images |
List[str] | up to 9 items | Up to 9 reference image URLs. At least one reference image or video is required. Images must be 256-5760 pixels per side, with a width-to-height ratio of 0.4-2.5. | ||
reference_videos |
List[str] | up to 3 items | Up to 3 reference video URLs. Clips are adjusted to 2-15 seconds each, with a combined maximum of 15 seconds; longer clips are trimmed and shorter clips are extended. Use MP4 or MOV, up to 50 MB per file, 256-5760 pixels per side, width-to-height ratio 0.4-2.5, and 23.976-60 FPS. At least one image or video is required. | ||
reference_audios |
List[str] | up to 3 items | Up to 3 optional WAV or MP3 reference audio URLs, up to 15 MB per file. Each clip must be 2-15 seconds, with a combined maximum of 15 seconds. Audio is not automatically trimmed and cannot be provided alone. | ||
aspect_ratio |
str | "16:9" |
21:9, 16:9, 4:3, 1:1, 3:4, 9:16 |
Output aspect ratio. | |
resolution |
str | "768p" |
768p |
Output video resolution. | |
duration |
int | 5 |
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Output video duration in seconds. |
minimax/h3-image/text-to-image — MiniMax H3 Image Text to Image generates images from text prompts at 1K, 2K and 4K, with aspect ratio selection and built-in prompt enhancement.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | The positive prompt for the generation. | ||
aspect_ratio |
str | "1:1" |
1:1, 1:2, 2:1, 1:3, 3:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 9:21, 21:9 |
The aspect ratio of the generated image. Defaults to 1:1. | |
resolution |
str | "1k" |
1k, 2k, 4k |
The output resolution tier. 1k is about 1 megapixel, 2k about 4 megapixels and 4k about 16 megapixels. 1k and 2k cost the same; 4k is the highest. | |
output_format |
str | "png" |
png, jpeg, webp |
The format of the output image. |
minimax/h3-image/edit — MiniMax H3 Image Edit edits and composes images from 1 to 16 reference images and a text instruction, with aspect ratio selection and 1K, 2K and 4K output.
| Parameter | Type | Required | Default | Allowed | Description |
|---|---|---|---|---|---|
prompt |
str | yes | The positive prompt for the generation. | ||
images |
List[str] | yes | 1–16 items | The images to edit, 1 to 16. Refer to them in the prompt as image 1, image 2 and so on, in the order given. JPG, PNG, WEBP, HEIC and HEIF are supported; each side must be 128 to 8192 px and the long side at most 4 times the short side. | |
aspect_ratio |
str | 1:1, 1:2, 2:1, 1:3, 3:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 9:21, 21:9 |
The aspect ratio of the generated image. If not specified, it is determined automatically from the input images and the prompt. | ||
resolution |
str | "1k" |
1k, 2k, 4k |
The output resolution tier. 1k is about 1 megapixel, 2k about 4 megapixels and 4k about 16 megapixels. 1k and 2k cost the same; 4k is the highest. | |
output_format |
str | "png" |
png, jpeg, webp |
The format of the output image. |
The wrapper is a thin layer over the REST API: submit, then poll for the result.
curl -s -X POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/minimax-h3/text-to-video" \
-H "Authorization: Bearer $WAVESPEED_API_KEY" -H "Content-Type: application/json" \
-d '{"prompt": "A slow cinematic dolly shot through a rain-soaked neon night market, reflections on the pavement"}'
# -> {"data": {"id": "<prediction_id>"}}
curl -s "https://api.wavespeed.ai/api/v3/predictions/<prediction_id>/result" \
-H "Authorization: Bearer $WAVESPEED_API_KEY"
# -> status: completed, outputs: ["<url>"]npm i -g @wavespeed/cli && wavespeed login
wavespeed run wavespeed-ai/minimax-h3/text-to-video -p "your prompt"- awesome-minimax-h3-api — prompt library and variant guide
- More model wrappers: Seedance 2.5 · Wan 3.0 · Kling 4.0 · Veo 3.1 · HappyHorse 1.1 · FLUX 3 · Seedream 5.5 · GPT Image 2.6 · Nano Banana 2.1
- SDKs and tools: Python · JavaScript · CLI · MCP server · ComfyUI nodes
- Full model catalog
- MiniMax H3 is a model by MiniMax. This repository is maintained by WaveSpeedAI and wraps the WaveSpeedAI hosted API for it.
- Function signatures and the tables above are generated from the live model schemas (last generated 2026-11-22). The model pages are the source of truth if an endpoint changes.
- Issues and pull requests are welcome. License: MIT.