Gate.AIBlogHailuo 02: Full Specifications, Pricing, API Integration, and Use Cases (2026)

    Hailuo 02: Full Specifications, Pricing, API Integration, and Use Cases (2026)

    Models

    What Is Hailuo 02?

    Hailuo 02 is a short-video AI generation model released by MiniMax on June 18, 2025. It supports text-to-video and image-to-video generation. It offers 6-second and 10-second video options, supports camera motion instructions, and has been priced by video resolution and duration starting in July 2026. MiniMax states that Hailuo 02 has been integrated into the Hailuo Video web app, mobile apps, and the API platform. The initial official release includes 768P 6 seconds, 768P 10 seconds, and 1080P 6 seconds.

    Hailuo 02 is a video generation model, not a chat-style large language model (LLM). As a result, it does not have a traditional Token context window for long-document reasoning. In practice, the input limit is the video prompt limit: MiniMax’s official documentation specifies that text prompts for video generation tasks support up to 2,000 characters. Gate.AI’s documentation also states that the Hailuo prompt limit for video tasks is 2,000 characters.

    This model works well for comparing short-form AI video tools, planning the cost of single generations, testing products or marketing concepts, and building teams’ automated video workflows. For a broader AI video generation workflow, Hailuo 02 should be evaluated based on task type, resolution, duration, input mode, and review requirements—not by Token price as the only standard.

    Key Parameters and Pricing for Hailuo 02

    Parameter Verified values
    Provider MiniMax (as of July 2026)
    Model family Hailuo video generation model (as of July 2026)
    Model type A short-video AI generation model for workflows that support text-to-video and image-to-video generation (as of July 2026)
    Release date June 18, 2025 (as of July 2026)
    Context window Not applicable for Token context windows; MiniMax documentation video-generation prompt limit is 2,000 characters (as of July 2026)
    Input pricing Not Token-based; billed by video generation task, duration, and resolution (as of July 2026)
    Cached input pricing Not applicable; as of July 2026, Hailuo 02 video generation does not record any prompt-caching billing feature
    Output pricing 512P 6 seconds: \$0.10; 512P 10 seconds: \$0.15; 768P 6 seconds: \$0.28; 768P 10 seconds: \$0.56; 1080P 6 seconds: \$0.49; 1080P 10 seconds: not listed (as of July 2026)
    Billing unit Billed per generated video, based on resolution and duration (as of July 2026)
    Modality support Supports text-to-video and image-to-video generation (as of July 2026)
    Supported input types Text prompts; image-to-video supports the first-frame image via a public URL or a Base64 data URL (as of July 2026)
    Supported output types Video files generated through an asynchronous video task workflow (as of July 2026)
    API access MiniMax video generation API; Gate.AI asynchronous video API, see the Gate.AI list and documentation (as of July 2026)
    Provider API model name MiniMax-Hailuo-02 (as of July 2026)
    Gate.AI model ID minimax/hailuo-02, see the Gate.AI list (as of July 2026)
    Availability Hailuo web, mobile, and API platforms; Gate.AI access, see the Gate.AI list and video API documentation (as of July 2026)
    Knowledge cutoff Not applicable for this video generation model; as of July 2026, official documentation does not specify a training-data cutoff time
    Rate limits MiniMax video plan documentation lists RPM limits by plan tier; as of July 2026, the official docs do not fully list on-demand rate limits for this model
    Fine-tuning support Not confirmed by official docs as of July 2026
    Streaming support As of July 2026, only the asynchronous video generation workflow is recorded; streaming generation is not applicable
    Batch API support Not confirmed by official docs as of July 2026
    Tool/function calling Not applicable for this video generation model as of July 2026
    Structured output/JSON mode Not a model capability; API responses use JSON task objects in the asynchronous workflow (as of July 2026)
    License/usage restrictions General platform terms apply; as of July 2026, official model documentation does not fully list model-specific commercial-use restrictions

    MiniMax currently categorizes Hailuo 02 as an older Hailuo video model, but it still lists MiniMax-Hailuo-02 in its text-to-video and image-to-video API documentation. This means Hailuo 02 remains useful for cost-focused workflows using 512P, 768P, and short 1080P generations. When teams care more about model behavior updates, the Hailuo 2.3 series may be a better fit.

    What Are the Advantages of Hailuo 02 in Real Production?

    Hailuo 02 can generate short videos from text prompts. It’s suitable for creative concept work, dynamic mood-board tests, marketing previews, product story showcases, and rough drafts for social videos. The production constraint is duration. Hailuo 02 is designed for short clips, not continuous long scenes. MiniMax’s official documentation clearly supports generation options of 6 seconds and 10 seconds, and resolution support varies by task type and duration.

    The model also supports image-to-video generation, using the input image as the first frame. This is ideal for teams that want to start from pre-approved reference visuals—such as product renders, environment settings, character frames, or style boards. MiniMax requires that image inputs meet format, size, and aspect ratio requirements.

    Hailuo 02 supports camera motion instructions at the prompt level. MiniMax’s docs list instructions such as truck, pan, push, pull, pedestal, tilt, zoom, shake, tracking shot, and static shot. These controls help creators describe the camera direction more precisely, but they should be treated as guidance rather than frame-level motion guarantees.

    Hailuo 02 is well-suited for iterative cost control. Teams can first generate low-cost 512P or 768P drafts, then use 1080P generation for the final candidates. MiniMax’s on-demand pricing table clearly shows cost differences across 512P, 768P, and 1080P tiers.

    What Input and Output Modalities Does Hailuo 02 Support?

    Modality Supported? Notes
    Text input Supported Supports text prompts for video generation. MiniMax documentation prompt limit: 2,000 characters.
    Image input Supported Image-to-video supports the first-frame image, sourced from a public URL or a Base64 data URL.
    Video output Supported Outputs are generated via an asynchronous video task workflow.
    Audio output Not confirmed Official documentation has not confirmed native synchronous audio generation capabilities for this model.
    Text output Not supported Hailuo 02 is not a chat, text completion, or reasoning model.
    Multimodal understanding Limited to generation inputs Hailuo 02 can use text and image inputs to generate videos, but it does not support general multimodal reasoning.

    What Are the Limitations of Hailuo 02?

    Hailuo 02 is limited to short-form video generation. MiniMax lists 512P 6 seconds and 10 seconds, 768P 6 seconds and 10 seconds, and a 1080P 6-second option in its documentation and pricing tables, but it does not list 1080P 10 seconds. This makes it suitable for short clips, not for completing long sequences independently.

    Resolution support varies depending on the workflow. MiniMax’s text-to-video documentation lists 768P and 1080P options, while its image-to-video documentation lists 512P, 768P, and 1080P. Teams should confirm the task type and resolution before production.

    Generated videos may have artifacts, non-continuous motion, identity drift, prompt mismatches, or visual continuity errors. These are common limitations of AI video generation. Unless MiniMax explicitly states otherwise, these are not guaranteed to be model-exclusive. For commercial use, brand-sensitive content, or identity-sensitive output, human review is still required.

    Operationally, video generation runs as an asynchronous workflow. Applications need to be designed around task submission, status polling, retry handling, download management, billing checks, and failure recovery. Gate.AI documentation records video task submission, job_id polling, completion status, and temporary download URLs for video outputs.

    Hailuo 02 is not suitable for legal, medical, financial, political, or security-critical review. Before publishing generated videos—especially when prompts or reference images involve protected people, brands, characters, or media content—ensure rights, brand, consent, and copyright reviews are completed.

    What Scenarios Is Hailuo 02 Best For?

    Application scenario Why it fits Key limitations
    Social short-video creativity 6–10 second outputs are great for fast creative exploration. Not for fully producing long videos; you’ll need post-editing.
    Product or marketing preview Image-to-video can start from reference frames, helping preserve the initial visual direction. Product details may change across frames.
    Dynamic storyboard testing Camera instructions make it easier to specify frame directions. Camera instructions are only motion references, not precise cinematography guarantees.
    Low-cost draft generation 512P and 768P tiers reduce early experimentation costs. Low resolution is not suitable for final publishing.
    High-definition short clip generation The official pricing table lists 1080P 6-second outputs. 1080P 10-second outputs are not listed.
    Creative generation workflows Text and image inputs let marketing, design, and content teams think more flexibly. Outputs still require review for accuracy, visual quality, and rights risk.

    Hailuo 02 is best for video generation scenarios that are short, directional, and cost-controlled. If you need deterministic continuity, synchronous audio, long durations, strict character consistency, or high-risk content without human review, it’s not recommended.

    Hailuo 02 vs. MiniMax Hailuo 2.3 and MiniMax Hailuo 2.3 Fast

    Comparison dimension Hailuo 02 MiniMax Hailuo 2.3 MiniMax Hailuo 2.3 Fast Notes on suitable scenarios
    Model positioning Older Hailuo video model Next-generation Hailuo generation model Next-generation, efficiency-focused Hailuo model Hailuo 02 fits older workflows and low-cost 512P scenarios. The new model fits quality- or efficiency-driven needs.
    Text-to-video Supported Supported Not listed in the official model introduction table Hailuo 02 and Hailuo 2.3 are better suited for text-prompt video generation.
    Image-to-video Supported Supported Supported All three support dynamic generation driven by the first frame or images.
    512P option Hailuo 02 has it Not listed in the official model introduction table Not listed in the official model introduction table Hailuo 02 is suitable for low-cost 512P draft generation.
    768P option 6 seconds and 10 seconds available 6 seconds and 10 seconds available 6 seconds and 10 seconds available All three support 768P short-form video generation.
    1080P option 6 seconds available 6 seconds available 6 seconds available All three are suitable for short high-definition outputs.
    Price \$0.10–\$0.56, billed by resolution and duration \$0.28–\$0.56, billed by resolution and duration \$0.19–\$0.33, billed by resolution and duration For cost-sensitive scenarios, compare the specific task, duration, and resolution.
    Best-fit scenario comparison Cost-focused short video generation with 512P support Updated Hailuo video quality path Efficient image-to-video generation Selection depends on your workflow—no absolute "better or worse."

    MiniMax Hailuo 2.3 is the most direct quality-oriented comparison because it’s the next-generation Hailuo model listed alongside Hailuo 02 in MiniMax’s docs. When the main need is efficient image-to-video generation and a lower price, Hailuo 2.3 Fast is a better choice.

    How to Access Hailuo 02 via Gate.AI?

    According to the Gate.AI list, the model ID for Hailuo 02 is minimax/hailuo-02. Gate.AI documentation verifies the video generation API flow, including POST /api/v1/videos, Bearer Token authentication, asynchronous task submission, job_id status polling, and downloading the video after completion. Gate.AI also notes that the Hailuo video prompt limit is 2,000 characters; if the prompt exceeds the model limit, it will be rejected upstream.

    Gate.AI’s pricing and documentation pages detail platform capabilities, including on-demand billing, model routing, API key management, usage insights, budget protection, organization permissions, and model access via the API protocol. These platform features should be distinguished from the provider-side model parameters and pricing on MiniMax’s side.

    Python example

    1. import os
    2. import requests
    3. api_key = os.environ["GATEAI_API_KEY"]
    4. payload = {
    5. "model": "minimax/hailuo-02",
    6. "prompt": "A close-up of a glass teacup on a rainy window ledge [Push in], cinematic lighting",
    7. "duration": 6,
    8. "resolution": "768p",
    9. "aspect_ratio": "16:9",
    10. "generate_audio": False
    11. }
    12. response = requests.post(
    13. "https://api.gate.ai/api/v1/videos",
    14. headers={
    15. "Authorization": f"Bearer {api_key}",
    16. "Content-Type": "application/json"
    17. },
    18. json=payload,
    19. timeout=60
    20. )
    21. response.raise_for_status()
    22. print(response.json())

    curl example

    1. curl https://api.gate.ai/api/v1/videos \
    2. -H "Authorization: Bearer $GATEAI_API_KEY" \
    3. -H "Content-Type: application/json" \
    4. -d '{
    5. "model": "minimax/hailuo-02",
    6. "prompt": "A close-up of a glass teacup on a rainy window ledge [Push in], cinematic lighting",
    7. "duration": 6,
    8. "resolution": "768p",
    9. "aspect_ratio": "16:9",
    10. "generate_audio": false
    11. }'

    The submission response returns an asynchronous task object, including fields such as job_id, status, model, and status_url. Your application should poll GET /api/v1/videos/{job_id} until completion, and then download the output video following the documentation flow.

    FAQ

    What resolutions does Hailuo 02 support?
    Hailuo 02 supports 512P, 768P, and 1080P based on task type and duration. MiniMax lists 6-second and 10-second options for 512P and 768P. As of July 2026, for 1080P it lists only 6-second outputs.

    How much does Hailuo 02 cost?
    MiniMax lists on-demand pricing for Hailuo 02: 512P 6 seconds \$0.10, 512P 10 seconds \$0.15, 768P 6 seconds \$0.28, 768P 10 seconds \$0.56, 1080P 6 seconds \$0.49 (as of July 2026).

    Can I access Hailuo 02 via API?
    Yes. MiniMax documentation records the MiniMax-Hailuo-02 video generation API. In Gate.AI’s list, the model ID is minimax/hailuo-02, used together with Gate.AI’s asynchronous video task API.

    What is Hailuo 02 best used for?
    Hailuo 02 is suitable for short creative videos, dynamic storyboard testing, marketing previews, product concept clips, and experiments with image-to-video generation. It’s not suitable for deterministic long-video production without editing, rights review, and manual quality checks.

    The content herein does not constitute any offer, solicitation, or recommendation. You should always seek independent professional advice before making any investment decisions. Please note that Gate may restrict or prohibit the use of all or a portion of the Services from Restricted Locations. For more information, please read the User Agreement

    Related Articles