Hunyuan Image v3 Instruct Edit

Edit images with natural language using intelligent reasoning. Supports up to 2 reference images.

Input

Original

Output

Generated

Upload your image and transform it in seconds

12,000+ images created this month

📄 About Hunyuan Image v3 Instruct Edit

Hunyuan Image v3 Instruct Edit is an advanced AI-powered image editing model developed by Tencent, designed to revolutionize the way users transform and enhance images. Leveraging intelligent internal reasoning capabilities, this model interprets natural language prompts to apply precise and creative edits, making complex image manipulations accessible to users of all skill levels. With support for up to two reference images, Hunyuan Image v3 Instruct Edit enables nuanced transformations, style transfers, and content blending, providing unparalleled flexibility for both professional and personal projects. The core technology behind this model combines state-of-the-art deep learning with sophisticated image processing algorithms. Users can input one or two images and describe the desired transformation in plain language—such as "turn this artwork into a realistic image"—and the model responds by creating high-quality, visually coherent output. The platform offers a range of customizable parameters, including output image size, number of images generated (from 1 to 4), guidance scale for prompt adherence, and output format (JPEG or PNG). This allows users to tailor results for their specific creative or professional needs, whether they are producing content for marketing campaigns, refining digital artwork, or generating variants for A/B testing. Hunyuan Image v3 Instruct Edit is ideal for a variety of applications. Designers can use it to quickly iterate on visual concepts, while content creators can transform or enhance photos for social media and branding. Its intelligent reasoning is particularly valuable for tasks that require creative reinterpretation or style adaptation, such as turning sketches into realistic images or merging different visual styles. The model’s safety checker ensures that generated content is appropriate, making it suitable for use in educational and professional environments. Key capabilities include the ability to handle up to two reference images for complex edits, a spectrum of standard and custom output sizes, and fine-tuned control over the fidelity of the transformation to the user’s prompt. The model delivers results in a matter of seconds, streamlining workflows and boosting productivity. By using a pay-as-you-go credit system, users have the flexibility to scale their usage according to project demands without long-term commitments. Hunyuan Image v3 Instruct Edit stands out for its blend of ease-of-use and powerful features, catering to designers, marketers, educators, and anyone seeking smart, AI-driven image editing. Its natural-language interface makes advanced edits accessible, while its technical robustness ensures consistent, high-quality results. Whether you’re looking to enhance artwork, update product images, or experiment with creative visuals, this model provides a reliable and innovative solution.

✨ Key Features

Supports up to two reference images for complex and nuanced image edits.

Natural language prompt input enables intuitive and precise image transformations.

Multiple output image sizes and aspect ratios, including automatic, square, portrait, and landscape formats.

Generates up to four images per request for rapid iteration and comparison.

Adjustable guidance scale allows control over how strictly the output follows the prompt.

Choice of JPEG or PNG output formats for flexible use in various workflows.

Built-in safety checker helps ensure that generated images meet content standards.

💡 Use Cases

⚡Transforming digital artwork into realistic images for portfolios or presentations.

⚡Enhancing or retouching product photos for e-commerce and marketing campaigns.

⚡Blending styles or features from two images to create new visual concepts.

⚡Generating multiple image variations for A/B testing or creative exploration.

⚡Converting sketches or rough designs into polished, high-quality visuals.

⚡Updating or modernizing legacy images for brand refreshes.

⚡Creating educational materials by adapting historical images or illustrations.

🎯 Best For

🎯 Professional designers, marketers, content creators, and anyone seeking advanced yet user-friendly AI-powered image editing.

👍 Pros

✓Intelligent internal reasoning enables sophisticated and context-aware edits.

✓Easy-to-use interface with natural language prompts lowers the barrier for non-technical users.

✓Flexible input and output options accommodate a wide range of creative and professional needs.

✓Fast image generation speeds support efficient workflows.

✓Integrated safety checker provides added assurance for content appropriateness.

⚠️ Considerations

△Limited to a maximum of two reference images per edit.

△Advanced customization may require some experimentation with settings.

△Output is dependent on prompt clarity and quality of input images.

📚 How to Use Hunyuan Image v3 Instruct Edit

Prepare one or two reference images you would like to edit or transform.

Upload your images to the platform using the image upload feature.

Enter a detailed text prompt describing the desired transformation or edit.

Select your preferred output image size, number of images, guidance scale, and output format.

Click to generate the images and wait a few seconds for the results.

Review and download your edited images for use in your projects.

💡 Pro Tips for Hunyuan Image v3 Instruct Edit

★

Use Two Images for Complex Edits When you need to blend styles or combine elements from different sources, upload two reference images instead of one. The model's intelligent reasoning works best when it can analyze multiple visual inputs. For example, use one image for composition and another for style reference. This dual-image capability sets it apart from single-image editors like OpenAI GPT Image 2 Edit, making it ideal for sophisticated transformations that require contextual understanding from multiple sources.

★

Write Specific Transformation Prompts Instead of vague instructions like "make it better," describe exactly what you want: "convert this watercolor painting into a photorealistic portrait with natural skin tones and studio lighting." The model's natural language processing excels with detailed, action-oriented prompts. Include target style, lighting conditions, and specific elements to change. Clear prompts reduce the need for multiple iterations and help the guidance scale parameter work more effectively, saving you credits and time on each generation.

★

Adjust Guidance Scale for Control The default guidance scale of 3.5 balances creativity with prompt adherence, but you can adjust it based on your needs. Lower values (1-2.5) give the model more creative freedom, useful for artistic interpretations. Higher values (5-10) enforce stricter adherence to your prompt, better for precise edits like product photography. Experiment with this parameter when results don't match expectations. Models like FLUX 2 Dev Edit offer similar controls but with different creative biases.

★

Generate Multiple Variations Efficiently Set num_images to 2-4 to generate several variations in one request, which is more credit-efficient than running separate generations. This approach is perfect for A/B testing marketing visuals or exploring different creative directions. Review all outputs before committing to additional edits. The batch generation feature helps you identify which guidance scale and prompt phrasing work best for your specific reference images, establishing a workflow pattern you can repeat for similar projects.

★

Match Output Size to Your Use Case Choose "auto" for the model to decide optimal dimensions based on your input images, or select specific aspect ratios for platform requirements. Use square formats for social media profiles, landscape 16:9 for presentations and YouTube thumbnails, and portrait 9:16 for Instagram Stories or mobile content. Selecting the right size upfront eliminates the need for post-processing crops or resizes. For professional headshots with specific dimensions, consider AI Headshot Generator as a specialized alternative.

★

Start With High-Quality Reference Images Upload clear, well-lit images with sharp focus and minimal compression artifacts. The model's output quality directly correlates with input image quality. Avoid heavily filtered or low-resolution sources, as these limit the model's ability to perform intelligent reasoning about visual details. For best results, use images with at least 1024px on the shortest side. If your source images are sketches or low-quality scans, the model can still work but may require more prompt guidance to achieve professional results.

Ready to try Hunyuan Image v3 Instruct Edit?

Get 10 free credits — no credit card required

Start Free →

Frequently Asked Questions

The model can handle a wide range of edits, including style transfers, realism enhancements, blending features from multiple images, and transforming artwork into different visual styles. Simply describe your desired result in the prompt for best outcomes.

You can upload up to two reference images per session. This allows for more complex edits, such as blending styles or content from two different sources.

Pricing varies by model and is based on a pay-as-you-go credit system. This means you only pay for the credits you use, giving you flexibility and control over your spending.

Yes, you can select from several preset sizes or allow the model to decide automatically. The output format can be set to either JPEG or PNG, depending on your project requirements.

The model includes a safety checker to help ensure that generated images meet content standards. However, it’s always recommended to review outputs for specific use cases to ensure compliance.

Pricing operates on a pay-as-you-go credit system that varies by model and generation parameters. Each edit consumes credits based on output resolution, number of images generated, and processing complexity. Generating four images in one request is typically more credit-efficient than four separate single-image requests. You can view exact credit costs before generating by checking the model's pricing details on JAI Portal. There are no subscription fees or monthly minimums—you only pay for what you use. New users often receive starter credits to test the model. For budget planning, start with single-image generations at default settings to establish baseline costs, then scale up as needed for production work.

Yes, all paid output generated on JAI Portal includes commercial-use rights, meaning you can use edited images in client work, marketing campaigns, product listings, and other revenue-generating projects. This applies to images created with Hunyuan Image v3 Instruct Edit as long as you've paid credits for the generation. However, you remain responsible for ensuring your input images don't violate third-party copyrights or trademarks. The built-in safety checker helps filter inappropriate content, but it's your responsibility to verify that reference images are either your original work, properly licensed, or in the public domain. For high-stakes commercial use, review outputs carefully and maintain records of your input sources and generation parameters.

Both models excel at natural language image editing but with different strengths. Hunyuan Image v3 Instruct Edit supports up to two reference images and uses intelligent internal reasoning for complex transformations, making it ideal for blending styles or combining visual elements from multiple sources. FLUX 2 Dev Edit typically focuses on single-image edits with different creative biases and may offer faster generation times for simpler tasks. Hunyuan's dual-image capability is unique among instruction-based editors on JAI Portal. If your workflow requires merging content from two images or you need sophisticated style transfers, Hunyuan is the better choice. For straightforward single-image edits or when you prefer FLUX's aesthetic, try both models with the same prompt to compare results.

Use the seed parameter to ensure reproducible results. When you generate an image you like, note the seed value from the output metadata. Enter that same seed in your next generation along with identical prompt text, guidance scale, and image size settings to recreate similar results. This is essential for maintaining visual consistency across a series of images or when you need to regenerate an output with minor prompt adjustments. Keep in mind that changing reference images or other parameters will alter results even with the same seed. For production workflows requiring batch consistency, document all generation parameters including seed values. This reproducibility feature is available across most JAI Portal editing models, making it easy to establish and maintain brand-consistent visual styles.

JAI Portal provides API access for integrating models like Hunyuan Image v3 Instruct Edit into your own applications, automation workflows, or content pipelines. The API uses the same pay-as-you-go credit system as the web interface, with credits deducted per generation. You can programmatically upload reference images, submit prompts, adjust parameters like guidance scale and output format, and retrieve generated images. This enables batch processing, automated content creation, and integration with tools like content management systems or marketing automation platforms. API documentation includes code examples and rate limit information. For developers building image editing features into SaaS products or agencies managing high-volume client work, API access streamlines operations and eliminates manual uploads. Check the JAI Portal developer documentation for authentication details and endpoint specifications.

⚖️ How Hunyuan Image v3 Instruct Edit Compares

Hunyuan Image v3 Instruct Edit occupies a unique position among JAI Portal's image editing models due to its dual-reference image capability and intelligent reasoning engine. Unlike single-image editors such as OpenAI GPT Image 2 Edit or FLUX 2 Dev Edit, Hunyuan can analyze and blend content from two source images simultaneously, making it the go-to choice for complex style transfers, content merging, and transformations that require contextual understanding from multiple visual inputs. This positions it between straightforward edit tools and specialized models like FLUX 2 Face to Full Portrait, which focuses exclusively on headshot-to-portrait expansion. For users who need instruction-based editing with more creative control than basic editors but don't require the specialized focus of niche models, Hunyuan offers an excellent middle ground. Its natural language interface makes sophisticated edits accessible without technical expertise, while adjustable parameters like guidance scale provide fine-tuned control for professionals. Compared to Qwen Image 2 Edit or Qwen Image 2 Pro Edit, Hunyuan's dual-image support and Tencent's training data give it distinct creative characteristics worth testing for your specific use case. The model works best when you need to combine visual concepts from multiple sources or perform transformations that benefit from reference-based reasoning. JAI Portal's side-by-side comparison tool lets you test Hunyuan against alternatives with the same prompt to find your ideal workflow. Sign up to experiment with starter credits across multiple editing models.