Can MiniMax H3 Max generate video from an image?

Yes. MiniMax H3 Max provides an image-to-video endpoint that uses an input image to guide video generation. Supplying an end image can also turn the workflow into first-to-last-frame generation.

Does MiniMax H3 Max generate synchronized audio?

Yes. H3 Max generates video with synchronized audio, allowing short scenes to include audiovisual elements rather than requiring a separate audio-generation stage.

How is MiniMax H3 Max different from MiniMax H3?

MiniMax H3 Max is a post-trained variant of MiniMax H3 developed by fal Research and optimized for prompt adherence, aesthetics, and inference throughput. Standard MiniMax H3 has a broader documented capability and resolution ceiling, including generation up to 2K.

How fast can MiniMax H3 Max generate an AI video?

fal reports that MiniMax H3 Max can generate a 5-second 768p video in about 2.78 seconds on its inference infrastructure. This is a reported provider-side latency result, not a guaranteed API response time for every workload.

What resolutions does the MiniMax H3 Max API support?

The core H3 Max API is documented around 480p and 768p video generation. This differs from standard MiniMax H3, whose documented generation workflow supports resolutions up to 2K.

Can MiniMax H3 Max handle detailed camera and motion instructions?

Yes. H3 Max was specifically post-trained for stronger prompt adherence, making detailed descriptions of subjects, actions, camera movement, environments, and visual style relevant to its intended use.

Is MiniMax H3 Max suitable for short-form social video generation?

Yes. Its 5–15 second generation range, portrait and landscape aspect ratios, synchronized audio, and low generation latency make it suitable for short-form video, advertising concepts, product clips, and rapid creative iteration.

MiniMax

Available

minimax-h3-max

MiniMax H3 Max API

MiniMax H3 Max is a high-end variant of MiniMax's H3 video-generation family. Documented H3 Max integrations support text-to-video and image-to-video generation. MiniMax's official H3 materials describe the broader H3 system as an omni-modal video model with native stereo audio.

Starting price

$0.064 / second

Usage-based billing

API health

Available

Production endpoint

Response

Async ready

Observed API response

Released

Aug 31, 2026

MiniMax

Playground

Try MiniMax H3 Max with a real request, review its parameters and validate the output before integrating the API.

Overview

A practical overview of MiniMax H3 Max, including important behavior and integration considerations.

Technical Specifications of MiniMax H3 Max

Specification	MiniMax H3 Max
Model ID	minimax-h3-max
Model family	MiniMax H3
Model type	Generative video model
Model origin	Post-trained from MiniMax H3 open weights
Post-training	fal Research
Primary optimization	Prompt adherence, aesthetics, inference throughput
Input	Text, image; reference-to-video support is available through the H3 Max family
Output	Video with synchronized/native audio
Duration	5–15 seconds
Resolution	480p and 768p; 1080p refinement availability depends on the current endpoint
Frame rate	24 FPS
Audio	Synchronized stereo audio
Text-to-video	Yes
Image-to-video	Yes
First-to-last frame	Supported through image-to-video with an end frame
Reference-to-video	Available through the H3 Max family/API
Aspect ratios	21:9, 16:9, 4:3, 1:1, 3:4, 9:16
Main API endpoints	api.cometapi.com/v1/videos
Base model	MiniMax H3

MiniMax H3 Max is a post-trained version of MiniMax H3 rather than a separately developed successor with a larger capability ceiling. fal Research states that it post-trained H3's open weights with additional data focused on prompt adherence and aesthetics, while its inference team co-designed the serving stack around the resulting model.

The distinction matters when comparing H3 Max with standard MiniMax H3. H3 itself is a general-purpose omni-modal video system supporting text, image, video, and audio understanding and video generation with native stereo audio. MiniMax's open-source H3 documentation specifies output durations of 4–15 seconds and resolutions up to 2K through its H3 variants.

H3 Max instead focuses on a different operating point: faster generation, strong prompt following, and visual quality at 480p/768p output. fal reports that a 5-second 768p clip can be generated in under three seconds on its infrastructure.

What is MiniMax H3 Max?

MiniMax H3 Max is an AI video generation model created through post-training of the open-weight MiniMax H3 model. fal Research describes the model as specifically tuned for stronger prompt adherence and aesthetics, while its inference stack is optimized for high throughput.

The model supports text-to-video and image-to-video generation, with the image-to-video endpoint also supporting first-to-last-frame generation when an end image is supplied. The H3 Max API family additionally exposes reference-to-video functionality.

Its most distinctive characteristic is therefore not a larger parameter count or a longer context window. It is the combination of video quality, prompt adherence, and low generation latency.

Main Features of MiniMax H3 Max

Post-trained MiniMax H3 foundation: H3 Max is derived from MiniMax H3's open weights rather than being an unrelated video model. fal Research added post-training data targeted at prompt adherence and aesthetics.
Fast video generation: fal reports generating a 5-second 768p clip in approximately 2.78 seconds on its infrastructure, substantially reducing the waiting time associated with iterative video generation.
Strong prompt adherence: The post-training process specifically targets better adherence to detailed instructions, including scene composition, actions, camera movement, and visual style.
Text-to-video and image-to-video: Developers can create video directly from a text prompt or use an image as the starting frame. The image endpoint can also accept an end frame for first-to-last-frame generation.
Native synchronized audio: H3 Max generates video with synchronized audio, making it suitable for short-form scenes that require dialogue, environmental sound, or audiovisual coordination.
Multiple aspect ratios: The model supports common landscape, portrait, square, and cinematic formats, including 21:9, 16:9, 4:3, 1:1, 3:4, and 9:16.

Benchmark Performance of MiniMax H3 Max

Evaluation	MiniMax H3 Max Result	Evaluation Type	Sample Size / Confidence	What It Measures
Design Arena – Image-to-Video	1,341 Elo	Human preference Elo	Arena-based comparison	Overall user preference for generated video quality
Artificial Analysis – Image-to-Video + Audio	1,201 Elo	Human preference Elo	2,177 samples; ±11 at 95% CI	Preference for image-to-video generation with audio
fal Human Preference Evaluation – Overall Quality	#1 among evaluated models	Head-to-head human evaluation	Compared with 12 leading video models	Overall visual quality
fal Human Preference Evaluation – Prompt Understanding	#1 among evaluated models	Head-to-head human evaluation	Compared with 12 leading video models	How accurately the generated video follows the prompt
fal Human Preference Evaluation – Aesthetics	#1 among evaluated models	Head-to-head human evaluation	Compared with 12 leading video models	Visual aesthetics and perceived quality
Generation Speed – 5-second 768p video	<3 seconds	Inference-speed measurement	fal infrastructure	End-to-end generation speed
Generation Speed vs Official MiniMax H3	~35× higher throughput	Provider-reported throughput comparison	fal infrastructure	Relative inference throughput

The strongest quantitative public results are the 1,341 Elo Design Arena score and the 1,201 Elo Artificial Analysis score. However, both are human-preference Elo measurements, not conventional accuracy benchmarks. The Artificial Analysis result reports a 95% confidence interval of ±11 over 2,177 samples.

For CometAPI content, the safest formulation is therefore to distinguish:

Capability evidence: prompt adherence, aesthetics, audiovisual generation and image/video conditioning.

Performance evidence: provider-reported generation latency and third-party preference evaluations.

Avoid presenting "#1" human-preference result as an objective overall leaderboard position.

MiniMax H3 Max vs MiniMax H3

Capability	MiniMax H3 Max	MiniMax H3
Origin	H3 open weights + fal post-training	Original MiniMax H3
Primary focus	Prompt adherence + aesthetics + speed	General-purpose omni-modal video generation
Typical output duration	5–15 sec	4–15 sec
Standard output	480p / 768p	Up to 2K through documented H3 variants
Native audio	Yes	Yes
Text-to-video	Yes	Yes
Image-to-video	Yes	Yes
First/last-frame workflow	Supported	Supported
Reference workflows	H3 Max reference endpoint available	Broad reference-to-video capabilities
Generation speed	Optimized for high throughput	Standard H3 inference
Best documented distinction	Fast iterative generation	Broader capability/resolution ceiling

The most important difference is operating point rather than model generation. Standard H3 is designed as a broader omni-modal system and supports output up to 2K in MiniMax's documentation, whereas H3 Max was developed around faster generation and prompt adherence at its supported output resolutions.

Representative Use Cases of MiniMax H3 Max

Rapid creative iteration

H3 Max is well suited to workflows where creators repeatedly generate, inspect, and revise short clips. Lower latency makes it practical to test several prompt variations rather than waiting several minutes for each iteration.

Its short-duration format, portrait aspect ratio support, synchronized audio, and fast generation make H3 Max suitable for short-form social content and advertising concepts.

Product visualization

Image-to-video generation can animate product stills into short promotional sequences while allowing the source image to guide composition and appearance.

Cinematic concept development

Detailed prompts describing camera movement, subject action, environment, lighting, and visual style can be used to prototype cinematic shots before committing to more expensive production workflows.

First-to-last-frame transitions

When an opening and ending image are supplied, H3 Max can be used for controlled transition sequences rather than relying entirely on unconstrained text-to-video generation.

How to Access MiniMax H3 Max API with CometAPI

CometAPI can be used as a unified API layer for integrating supported AI models without maintaining separate provider-specific integrations for every model.

Step 1: Create a CometAPI API key

Step 2: Select `minimax-h3-max`

Use minimax-h3-max as the model identifier when the model is available in your CometAPI account.

Step 3: Send a video-generation request

Submit your prompt and the supported generation parameters through CometAPI's video API interface. Depending on the currently exposed CometAPI schema, parameters may include duration, resolution, aspect ratio, image inputs, and other video-generation controls.

Before publishing an integration example, verify the current CometAPI model page and API documentation for the exact endpoint, parameter names, supported resolutions, and asynchronous task-handling behavior. These details can change independently of the underlying H3 Max model.

FAQ

Pricing

Use the quick estimate above for a single run, then review the full CometAPI and official-price comparison.

5-second test

$0.32

Estimated CometAPI usage cost

10-second test

$0.64

Estimated CometAPI usage cost

15-second test

$0.96

Estimated CometAPI usage cost

Pricing for MiniMax H3 Max

Explore competitive pricing for MiniMax H3 Max, designed to fit various budgets and usage needs. Our flexible plans ensure you only pay for what you use, making it easy to scale as your requirements grow. Discover how MiniMax H3 Max can enhance your projects while keeping costs manageable.

Comet Price (USD / M Tokens)	Official Price (USD / M Tokens)	Discount
Per Second:$0.064	Per Second:$0.08	-20%

API

Copy a working endpoint and code example, then open the complete API reference when you need every parameter.

One API, production-ready model access

Authenticate once, call the model endpoint and keep the same billing and observability workflow across providers.

Sample code and API for MiniMax H3 Max

Access comprehensive sample code and API resources for MiniMax H3 Max to streamline your integration process. Our detailed documentation provides step-by-step guidance, helping you leverage the full potential of MiniMax H3 Max in your projects.

set -euo pipefail

# 1. Create the task.
task=$(curl --fail-with-body --silent --show-error \
  https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  --form-string 'model=minimax-h3-max' \
  --form-string 'prompt=A cinematic view of clouds moving over green mountains, locked camera, no text.' \
  --form-string 'seconds=5' \
  --form-string 'size=1344x768')
# Add --form-string 'images=https://example.com/reference.jpg' for image-to-video.

task_id=$(jq -r '.id' <<<"$task")
echo "Task: ${task_id}"

# 2. Poll until the task finishes.
while true; do
  task=$(curl --fail-with-body --silent --show-error \
    --retry 3 --retry-all-errors --retry-delay 2 \
    "https://api.cometapi.com/v1/videos/${task_id}" \
    -H "Authorization: Bearer $COMETAPI_KEY")
  status=$(jq -r '.status' <<<"$task")
  echo "Status: $status"
  [[ "$status" == "failed" ]] && { echo "$task" >&2; exit 1; }
  [[ "$status" == "completed" ]] && break
  sleep 10
done

# 3. Download the completed video.
curl --fail-with-body --silent --show-error \
  --retry 3 --retry-all-errors --retry-delay 2 \
  "https://api.cometapi.com/v1/videos/${task_id}/content" \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  --output "${task_id}.mp4"
echo "Saved: ${task_id}.mp4"

cURL Code Example

set -euo pipefail

# 1. Create the task.
task=$(curl --fail-with-body --silent --show-error \
  https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  --form-string 'model=minimax-h3-max' \
  --form-string 'prompt=A cinematic view of clouds moving over green mountains, locked camera, no text.' \
  --form-string 'seconds=5' \
  --form-string 'size=1344x768')
# Add --form-string 'images=https://example.com/reference.jpg' for image-to-video.

task_id=$(jq -r '.id' <<<"$task")
echo "Task: ${task_id}"

# 2. Poll until the task finishes.
while true; do
  task=$(curl --fail-with-body --silent --show-error \
    --retry 3 --retry-all-errors --retry-delay 2 \
    "https://api.cometapi.com/v1/videos/${task_id}" \
    -H "Authorization: Bearer $COMETAPI_KEY")
  status=$(jq -r '.status' <<<"$task")
  echo "Status: $status"
  [[ "$status" == "failed" ]] && { echo "$task" >&2; exit 1; }
  [[ "$status" == "completed" ]] && break
  sleep 10
done

# 3. Download the completed video.
curl --fail-with-body --silent --show-error \
  --retry 3 --retry-all-errors --retry-delay 2 \
  "https://api.cometapi.com/v1/videos/${task_id}/content" \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  --output "${task_id}.mp4"
echo "Saved: ${task_id}.mp4"

Python Code Example

import os
import time
from pathlib import Path

import requests


api_key = os.environ["COMETAPI_KEY"]
base_url = "https://api.cometapi.com"
headers = {"Authorization": f"Bearer {api_key}"}


def get(path: str, timeout: int) -> requests.Response:
    for attempt in range(3):
        try:
            response = requests.get(
                f"{base_url}{path}", headers=headers, timeout=timeout
            )
            response.raise_for_status()
            return response
        except requests.RequestException:
            if attempt == 2:
                raise
            time.sleep(2)
    raise RuntimeError("unreachable")

# 1. Create the task.
form = {
    "model": (None, "minimax-h3-max"),
    "prompt": (
        None,
        "A cinematic view of clouds moving over green mountains, "
        "locked camera, no text.",
    ),
    "seconds": (None, "5"),
    "size": (None, "1344x768"),
}
if image_url := os.getenv("MINIMAX_H3_IMAGE_URL"):
    form["images"] = (None, image_url)

response = requests.post(
    f"{base_url}/v1/videos",
    headers=headers,
    files=form,
    timeout=60,
)
response.raise_for_status()
task_id = response.json()["id"]
print(f"Task: {task_id}")

# 2. Poll until the task finishes.
while True:
    response = get(f"/v1/videos/{task_id}", timeout=30)
    task = response.json()
    status = task["status"]
    print(f"Status: {status}")
    if status == "failed":
        raise RuntimeError(task)
    if status == "completed":
        break
    time.sleep(10)

# 3. Download the completed video.
response = get(f"/v1/videos/{task_id}/content", timeout=300)
output = Path(f"{task_id}.mp4")
output.write_bytes(response.content)
print(f"Saved: {output}")

JavaScript Code Example

import fs from "node:fs";

const apiKey = process.env.COMETAPI_KEY;
const baseUrl = "https://api.cometapi.com";
const headers = { Authorization: `Bearer ${apiKey}` };
const sleep = (ms) => new Promise((resolve) => setTimeout(resolve, ms));

async function get(path) {
  for (let attempt = 0; attempt < 3; attempt += 1) {
    try {
      const result = await fetch(`${baseUrl}${path}`, { headers });
      if (!result.ok) throw new Error(`${result.status} ${await result.text()}`);
      return result;
    } catch (error) {
      if (attempt === 2) throw error;
      await sleep(2000);
    }
  }
  throw new Error("unreachable");
}

// 1. Create the task.
const form = new FormData();
form.append("model", "minimax-h3-max");
form.append(
  "prompt",
  "A cinematic view of clouds moving over green mountains, locked camera, no text.",
);
form.append("seconds", "5");
form.append("size", "1344x768");
if (process.env.MINIMAX_H3_IMAGE_URL) {
  form.append("images", process.env.MINIMAX_H3_IMAGE_URL);
}

let response = await fetch(`${baseUrl}/v1/videos`, {
  method: "POST",
  headers,
  body: form,
});
if (!response.ok) throw new Error(`${response.status} ${await response.text()}`);
const taskId = (await response.json()).id;
console.log(`Task: ${taskId}`);

// 2. Poll until the task finishes.
while (true) {
  response = await get(`/v1/videos/${taskId}`);
  const task = await response.json();
  console.log(`Status: ${task.status}`);
  if (task.status === "failed") throw new Error(JSON.stringify(task));
  if (task.status === "completed") break;
  await sleep(10000);
}

// 3. Download the completed video.
response = await get(`/v1/videos/${taskId}/content`);
const output = `${taskId}.mp4`;
fs.writeFileSync(output, Buffer.from(await response.arrayBuffer()));
console.log(`Saved: ${output}`);

Capabilities

Scan the model facts that matter before you choose an architecture or estimate production workload.

Supported capabilities

Features returned by the existing model API.

text-to-video

image-to-video

video-editing

Input and output

Inputs

text

Outputs

video

Production profile

Provider: MiniMax
Context: Model specific
Commercial use: Supported

Use cases

Use MiniMax H3 Max for production workflows that match its video capabilities, then compare alternatives before committing to a long-term integration.

Text-to-video concepts and social content

Product demos and cinematic storyboards

Image-to-video motion and creative iteration

Other models

Compare other models available through CometAPI for different quality, latency, capability and pricing trade-offs.

Open model comparison →

MiniMax

MiniMax H3

minimax-h3 is a new video generation model designed for high-quality creative production. It delivers improved prompt understanding, stronger visual consistency, smoother motion, and more flexible reference-image control for cinematic video generation workflows.

View model

MiniMax

minimax_video

Explore the minimax_video API.

View model

Aliyun

Wan3.0

From fast verification to cinematic output, billing is based on the generated video duration, selecting the appropriate resolution, and turning every generation into a planned creative investment activity rule

View model

Flux

Flux 3

FLUX 3 - Real World Models: Towards Multimodal Flow Models as the Backbone of Visual Intelligence.

View model

Aliyun

Happy Horse 1.1

HappyHorse 1.1 is a multimodal video-generation model designed for professional content creation, advertising, short films, social media production, and storytelling. It extends the capabilities of HappyHorse 1.0—which gained significant attention after ranking highly in independent video-generation evaluations—with stronger scene coherence and improved visual fidelity.

View model

Bytedance

Seedance 2-0

Seedance 2.0 is ByteDance’s next-generation multimodal video foundation model focused on cinematic, multi-shot narrative video generation. Unlike single-shot text-to-video demos, Seedance 2.0 emphasizes reference-based control (images, short clips, audio), coherent character/style consistency across shots, and native audio/video synchronization — aiming to make AI video useful for professional creative and previsualization workflows.

View model

Uptime

Review live heartbeat data, endpoint availability and observed response times before moving into production.

Keep exploring

Compare models and continue your research

MiniMax

Available

minimax-h3-max

MiniMax H3 Max API

Starting price

$0.064 / second

Usage-based billing

API health

Available

Production endpoint

Response

Async ready

Observed API response

Released

Aug 31, 2026

MiniMax

Playground

Try MiniMax H3 Max with a real request, review its parameters and validate the output before integrating the API.

Overview

A practical overview of MiniMax H3 Max, including important behavior and integration considerations.

Technical Specifications of MiniMax H3 Max

Specification	MiniMax H3 Max
Model ID	minimax-h3-max
Model family	MiniMax H3
Model type	Generative video model
Model origin	Post-trained from MiniMax H3 open weights
Post-training	fal Research
Primary optimization	Prompt adherence, aesthetics, inference throughput
Input	Text, image; reference-to-video support is available through the H3 Max family
Output	Video with synchronized/native audio
Duration	5–15 seconds
Resolution	480p and 768p; 1080p refinement availability depends on the current endpoint
Frame rate	24 FPS
Audio	Synchronized stereo audio
Text-to-video	Yes
Image-to-video	Yes
First-to-last frame	Supported through image-to-video with an end frame
Reference-to-video	Available through the H3 Max family/API
Aspect ratios	21:9, 16:9, 4:3, 1:1, 3:4, 9:16
Main API endpoints	api.cometapi.com/v1/videos
Base model	MiniMax H3

What is MiniMax H3 Max?

Its most distinctive characteristic is therefore not a larger parameter count or a longer context window. It is the combination of video quality, prompt adherence, and low generation latency.

Main Features of MiniMax H3 Max

Post-trained MiniMax H3 foundation: H3 Max is derived from MiniMax H3's open weights rather than being an unrelated video model. fal Research added post-training data targeted at prompt adherence and aesthetics.
Fast video generation: fal reports generating a 5-second 768p clip in approximately 2.78 seconds on its infrastructure, substantially reducing the waiting time associated with iterative video generation.
Strong prompt adherence: The post-training process specifically targets better adherence to detailed instructions, including scene composition, actions, camera movement, and visual style.
Text-to-video and image-to-video: Developers can create video directly from a text prompt or use an image as the starting frame. The image endpoint can also accept an end frame for first-to-last-frame generation.
Native synchronized audio: H3 Max generates video with synchronized audio, making it suitable for short-form scenes that require dialogue, environmental sound, or audiovisual coordination.
Multiple aspect ratios: The model supports common landscape, portrait, square, and cinematic formats, including 21:9, 16:9, 4:3, 1:1, 3:4, and 9:16.

Benchmark Performance of MiniMax H3 Max

Evaluation	MiniMax H3 Max Result	Evaluation Type	Sample Size / Confidence	What It Measures
Design Arena – Image-to-Video	1,341 Elo	Human preference Elo	Arena-based comparison	Overall user preference for generated video quality
Artificial Analysis – Image-to-Video + Audio	1,201 Elo	Human preference Elo	2,177 samples; ±11 at 95% CI	Preference for image-to-video generation with audio
fal Human Preference Evaluation – Overall Quality	#1 among evaluated models	Head-to-head human evaluation	Compared with 12 leading video models	Overall visual quality
fal Human Preference Evaluation – Prompt Understanding	#1 among evaluated models	Head-to-head human evaluation	Compared with 12 leading video models	How accurately the generated video follows the prompt
fal Human Preference Evaluation – Aesthetics	#1 among evaluated models	Head-to-head human evaluation	Compared with 12 leading video models	Visual aesthetics and perceived quality
Generation Speed – 5-second 768p video	<3 seconds	Inference-speed measurement	fal infrastructure	End-to-end generation speed
Generation Speed vs Official MiniMax H3	~35× higher throughput	Provider-reported throughput comparison	fal infrastructure	Relative inference throughput

For CometAPI content, the safest formulation is therefore to distinguish:

Capability evidence: prompt adherence, aesthetics, audiovisual generation and image/video conditioning.

Performance evidence: provider-reported generation latency and third-party preference evaluations.

Avoid presenting "#1" human-preference result as an objective overall leaderboard position.

MiniMax H3 Max vs MiniMax H3

Capability	MiniMax H3 Max	MiniMax H3
Origin	H3 open weights + fal post-training	Original MiniMax H3
Primary focus	Prompt adherence + aesthetics + speed	General-purpose omni-modal video generation
Typical output duration	5–15 sec	4–15 sec
Standard output	480p / 768p	Up to 2K through documented H3 variants
Native audio	Yes	Yes
Text-to-video	Yes	Yes
Image-to-video	Yes	Yes
First/last-frame workflow	Supported	Supported
Reference workflows	H3 Max reference endpoint available	Broad reference-to-video capabilities
Generation speed	Optimized for high throughput	Standard H3 inference
Best documented distinction	Fast iterative generation	Broader capability/resolution ceiling

Representative Use Cases of MiniMax H3 Max

Rapid creative iteration

Its short-duration format, portrait aspect ratio support, synchronized audio, and fast generation make H3 Max suitable for short-form social content and advertising concepts.

Product visualization

Image-to-video generation can animate product stills into short promotional sequences while allowing the source image to guide composition and appearance.

Cinematic concept development

Detailed prompts describing camera movement, subject action, environment, lighting, and visual style can be used to prototype cinematic shots before committing to more expensive production workflows.

First-to-last-frame transitions

When an opening and ending image are supplied, H3 Max can be used for controlled transition sequences rather than relying entirely on unconstrained text-to-video generation.

How to Access MiniMax H3 Max API with CometAPI

CometAPI can be used as a unified API layer for integrating supported AI models without maintaining separate provider-specific integrations for every model.

Step 1: Create a CometAPI API key

Step 2: Select `minimax-h3-max`

Use minimax-h3-max as the model identifier when the model is available in your CometAPI account.

Step 3: Send a video-generation request

FAQ

Pricing

Use the quick estimate above for a single run, then review the full CometAPI and official-price comparison.

5-second test

$0.32

Estimated CometAPI usage cost

10-second test

$0.64

Estimated CometAPI usage cost

15-second test

$0.96

Estimated CometAPI usage cost

Pricing for MiniMax H3 Max

Comet Price (USD / M Tokens)	Official Price (USD / M Tokens)	Discount
Per Second:$0.064	Per Second:$0.08	-20%

API

Copy a working endpoint and code example, then open the complete API reference when you need every parameter.

One API, production-ready model access

Authenticate once, call the model endpoint and keep the same billing and observability workflow across providers.

Sample code and API for MiniMax H3 Max

set -euo pipefail

# 1. Create the task.
task=$(curl --fail-with-body --silent --show-error \
  https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  --form-string 'model=minimax-h3-max' \
  --form-string 'prompt=A cinematic view of clouds moving over green mountains, locked camera, no text.' \
  --form-string 'seconds=5' \
  --form-string 'size=1344x768')
# Add --form-string 'images=https://example.com/reference.jpg' for image-to-video.

task_id=$(jq -r '.id' <<<"$task")
echo "Task: ${task_id}"

# 2. Poll until the task finishes.
while true; do
  task=$(curl --fail-with-body --silent --show-error \
    --retry 3 --retry-all-errors --retry-delay 2 \
    "https://api.cometapi.com/v1/videos/${task_id}" \
    -H "Authorization: Bearer $COMETAPI_KEY")
  status=$(jq -r '.status' <<<"$task")
  echo "Status: $status"
  [[ "$status" == "failed" ]] && { echo "$task" >&2; exit 1; }
  [[ "$status" == "completed" ]] && break
  sleep 10
done

# 3. Download the completed video.
curl --fail-with-body --silent --show-error \
  --retry 3 --retry-all-errors --retry-delay 2 \
  "https://api.cometapi.com/v1/videos/${task_id}/content" \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  --output "${task_id}.mp4"
echo "Saved: ${task_id}.mp4"

cURL Code Example

set -euo pipefail

# 1. Create the task.
task=$(curl --fail-with-body --silent --show-error \
  https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  --form-string 'model=minimax-h3-max' \
  --form-string 'prompt=A cinematic view of clouds moving over green mountains, locked camera, no text.' \
  --form-string 'seconds=5' \
  --form-string 'size=1344x768')
# Add --form-string 'images=https://example.com/reference.jpg' for image-to-video.

task_id=$(jq -r '.id' <<<"$task")
echo "Task: ${task_id}"

# 2. Poll until the task finishes.
while true; do
  task=$(curl --fail-with-body --silent --show-error \
    --retry 3 --retry-all-errors --retry-delay 2 \
    "https://api.cometapi.com/v1/videos/${task_id}" \
    -H "Authorization: Bearer $COMETAPI_KEY")
  status=$(jq -r '.status' <<<"$task")
  echo "Status: $status"
  [[ "$status" == "failed" ]] && { echo "$task" >&2; exit 1; }
  [[ "$status" == "completed" ]] && break
  sleep 10
done

# 3. Download the completed video.
curl --fail-with-body --silent --show-error \
  --retry 3 --retry-all-errors --retry-delay 2 \
  "https://api.cometapi.com/v1/videos/${task_id}/content" \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  --output "${task_id}.mp4"
echo "Saved: ${task_id}.mp4"

Python Code Example

import os
import time
from pathlib import Path

import requests


api_key = os.environ["COMETAPI_KEY"]
base_url = "https://api.cometapi.com"
headers = {"Authorization": f"Bearer {api_key}"}


def get(path: str, timeout: int) -> requests.Response:
    for attempt in range(3):
        try:
            response = requests.get(
                f"{base_url}{path}", headers=headers, timeout=timeout
            )
            response.raise_for_status()
            return response
        except requests.RequestException:
            if attempt == 2:
                raise
            time.sleep(2)
    raise RuntimeError("unreachable")

# 1. Create the task.
form = {
    "model": (None, "minimax-h3-max"),
    "prompt": (
        None,
        "A cinematic view of clouds moving over green mountains, "
        "locked camera, no text.",
    ),
    "seconds": (None, "5"),
    "size": (None, "1344x768"),
}
if image_url := os.getenv("MINIMAX_H3_IMAGE_URL"):
    form["images"] = (None, image_url)

response = requests.post(
    f"{base_url}/v1/videos",
    headers=headers,
    files=form,
    timeout=60,
)
response.raise_for_status()
task_id = response.json()["id"]
print(f"Task: {task_id}")

# 2. Poll until the task finishes.
while True:
    response = get(f"/v1/videos/{task_id}", timeout=30)
    task = response.json()
    status = task["status"]
    print(f"Status: {status}")
    if status == "failed":
        raise RuntimeError(task)
    if status == "completed":
        break
    time.sleep(10)

# 3. Download the completed video.
response = get(f"/v1/videos/{task_id}/content", timeout=300)
output = Path(f"{task_id}.mp4")
output.write_bytes(response.content)
print(f"Saved: {output}")

JavaScript Code Example

import fs from "node:fs";

const apiKey = process.env.COMETAPI_KEY;
const baseUrl = "https://api.cometapi.com";
const headers = { Authorization: `Bearer ${apiKey}` };
const sleep = (ms) => new Promise((resolve) => setTimeout(resolve, ms));

async function get(path) {
  for (let attempt = 0; attempt < 3; attempt += 1) {
    try {
      const result = await fetch(`${baseUrl}${path}`, { headers });
      if (!result.ok) throw new Error(`${result.status} ${await result.text()}`);
      return result;
    } catch (error) {
      if (attempt === 2) throw error;
      await sleep(2000);
    }
  }
  throw new Error("unreachable");
}

// 1. Create the task.
const form = new FormData();
form.append("model", "minimax-h3-max");
form.append(
  "prompt",
  "A cinematic view of clouds moving over green mountains, locked camera, no text.",
);
form.append("seconds", "5");
form.append("size", "1344x768");
if (process.env.MINIMAX_H3_IMAGE_URL) {
  form.append("images", process.env.MINIMAX_H3_IMAGE_URL);
}

let response = await fetch(`${baseUrl}/v1/videos`, {
  method: "POST",
  headers,
  body: form,
});
if (!response.ok) throw new Error(`${response.status} ${await response.text()}`);
const taskId = (await response.json()).id;
console.log(`Task: ${taskId}`);

// 2. Poll until the task finishes.
while (true) {
  response = await get(`/v1/videos/${taskId}`);
  const task = await response.json();
  console.log(`Status: ${task.status}`);
  if (task.status === "failed") throw new Error(JSON.stringify(task));
  if (task.status === "completed") break;
  await sleep(10000);
}

// 3. Download the completed video.
response = await get(`/v1/videos/${taskId}/content`);
const output = `${taskId}.mp4`;
fs.writeFileSync(output, Buffer.from(await response.arrayBuffer()));
console.log(`Saved: ${output}`);

Capabilities

Scan the model facts that matter before you choose an architecture or estimate production workload.

Supported capabilities

Features returned by the existing model API.

text-to-video

image-to-video

video-editing

Input and output

Inputs

text

Outputs

video

Production profile

Provider: MiniMax
Context: Model specific
Commercial use: Supported

Use cases

Use MiniMax H3 Max for production workflows that match its video capabilities, then compare alternatives before committing to a long-term integration.

Text-to-video concepts and social content

Product demos and cinematic storyboards

Image-to-video motion and creative iteration