Precise Image Editing
Ask for a specific change without intentionally rebuilding every part of the image. This is useful for adjusting clothing, products, backgrounds, objects, or small visual details.
GPT Image 2.5 online
GPT Image 2.5 AI Image Generator & Editor
Create and edit images with GPT Image 2.5. Use Flare for faster generation or Sunburst for more precise, detail-focused creative work.
Model overview
What Is GPT Image 2.5?
GPT Image 2.5 is an OpenAI image model designed for image generation and image editing. On PWZ, it can create images from text prompts and apply targeted changes using up to 10 reference images.
The model family includes Flare and Sunburst. Flare focuses on faster generation for everyday creative work, while Sunburst spends longer on requests where added precision and fine detail matter.
It is useful for creators, marketing teams, ecommerce brands, and designers who need product visuals, ads, social content, concept images, or targeted image edits.
Examples
See What GPT Image 2.5 Can Create
Explore six GPT Image 2.5 outputs across different creative directions, compositions, and visual styles.
Features
GPT Image 2.5 Features
Generate, edit, and refine images with more control over references, composition, lighting, texture, style, and transparent-background workflows.
Reference Image Guidance
Use reference images to guide recognizable subjects, products, scenes, styles, and compositions.
Lighting and Texture Direction
Describe lighting, surfaces, materials, and texture details directly in the prompt.
Flexible Output Controls
Choose the aspect ratio, quality, image count, background, output format, and moderation setting.
Style Direction
Describe a visual style and composition for each generation or image-editing request.
Complex Image Composition
Combine subjects, layout direction, text, environment, lighting, and visual style in one detailed request.
Transparent Background Support
Choose a transparent background on PWZ and use PNG or WebP to preserve transparency.
Compare models
GPT Image 2.5 Flare vs Sunburst
Start with Flare when speed matters. Try Sunburst when the image needs more careful detail or precision.
Flare
Built for drafts, everyday creative work, social assets, and workflows where you want to test ideas quickly.
Sunburst
Spends longer on detailed final assets, careful edits, and work where fine detail is worth extra generation time.
AreaFlareSunburst
Main focusFaster generationAdded precision and fine detail
Best suited forDrafts, iteration, social assets, everyday creative workDetailed final assets, careful edits, precision-focused work
Generation timeFasterLonger
Image generationSupportedSupported
Image editingSupportedSupported
Reference-image workflowsSupportedSupported
Which should I start with?Start here for most iterative workflowsUse when extra detail is worth additional generation time
Same prompt test
Flare vs Sunburst Benchmark
Compare Flare and Sunburst outputs side by side using the same prompt and generation setup.
Test prompt
Premium editorial product photograph of a mechanical watch on dark slate, engraved dial details, realistic brushed metal, controlled studio reflections, shallow depth of field.
As the comparison shows, Flare produces striking results in less time, while Sunburst handles fine details more precisely, with improved texture and clarity. Although each model has its own strengths, Flare and Sunburst cost the same.
How to use
How to Use GPT Image 2.5
Move from a clear creative brief to a controlled edit in three steps.
- 01
Write Your Prompt
Describe the subject, setting, composition, lighting, style, and any text. For edits, state what should change and what should stay the same.
- 02
Add a Reference and Choose Your Model
Upload a reference when you need to preserve a subject, product, composition, or visual direction. Choose Flare for faster iteration or Sunburst for more detail and precision.
- 03
Generate and Refine
Choose the available output settings, generate your image, then make one focused change at a time for a more controlled workflow.
Prompt library
GPT Image 2.5 Prompt Examples
Copy a complete prompt, then adjust the subject, framing, lighting, text, and constraints for your own brief.
Minimal studio photograph of a white ceramic coffee maker, three-quarter view, soft morning window light, light gray background, subtle realistic shadows, clean ecommerce composition.
Vertical sports drink campaign image, cyclist climbing a mountain road at sunrise, dramatic perspective, clean space in the upper left for campaign copy, realistic commercial photography.
Overhead editorial photograph of handmade ramen, dark ceramic bowl, textured wooden table, warm side lighting, subtle steam, premium restaurant campaign photography.
Contemporary music poster titled “AFTER DARK”, date “NOV 14”, minimal black and cream palette, experimental Swiss-inspired grid, abstract architectural photography.
Clean 3D visual for a productivity app launch, floating translucent interface cards, soft studio lighting, neutral background, restrained modern technology aesthetic.
Replace the background with warm beige plaster. Keep the bottle, label, camera angle, reflections, and proportions unchanged.
Image editing
GPT Image 2.5 Image Editing
Upload reference images, describe a targeted change, and state clearly what should remain unchanged.
Change One Element
Original: Existing product photo
“Change the table from marble to dark walnut. Keep the product and lighting unchanged.”
Update a Background
Original: Portrait or product image
“Move the subject to a bright modern studio. Preserve the subject’s appearance and camera framing.”
Use Multiple References
Original image + supporting references
Upload up to 10 images to guide the subject, product, composition, or visual direction in one editing request.
Specifications
GPT Image 2.5 Technical Specifications
Reference values for the GPT Image 2.5 model family and its available generation controls.
- Model variants
- Flare and Sunburst
- Core workflows
- Text-to-image and image editing
- Aspect ratios
- Auto, 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16
- Quality options
- Auto, Low, Medium, High, Xhigh, Max
- Default quality
- Auto
- Images per generation
- 1 to 10
- Reference images
- Up to 10 images per edit request
- Background options
- Auto, Transparent, Opaque
- Output formats
- PNG, WebP, JPEG
- Moderation options
- Auto, Low
Choose Flare for speed or Sunburst for added detail and precision.
Create new images from prompts or edit existing reference images.
Choose a square, portrait, landscape, or automatic composition.
Balance generation cost, speed, and output detail for each request.
The preselected quality level for a new generation.
Generate between one and ten output images in a single task.
Guide subjects, products, composition, or visual direction with references.
Choose an automatic, transparent, or opaque background.
Download generated images as PNG, WebP, or JPEG files.
Choose the Auto or Low moderation setting for a generation.
Creative workflows
Built for Real Creative Work
Workflow 1
Content Creators
Create visuals for Instagram, TikTok, YouTube, thumbnails, personal projects, music artwork, fashion concepts, visual storytelling, and social campaigns. Use fast generations for ideation, then refine promising concepts.
Workflow 2
Marketing & Ad Teams
Produce campaign concepts, product reveals, paid-social visuals, display creative, posters, event graphics, and early-stage advertising directions without starting every variation from scratch.
Workflow 3
E-commerce & Brands
Create product scenes, ecommerce visuals, launch graphics, lifestyle concepts, marketplace imagery, and campaign assets while using reference inputs to help preserve product appearance.
Workflow 4
Design & Creative Teams
Explore compositions, campaign directions, moodboards, presentation visuals, concept art, interface marketing assets, and visual prototypes with Flare or detail-focused Sunburst workflows.
Limitations
What GPT Image 2.5 May Still Get Wrong
AI image generation is not deterministic, and a strong prompt does not guarantee an exact result. Review production images before publishing.
Typography and spelling
Review small or dense typography and exact spelling inside complex layouts.
Product and brand details
Compare fine product details and exact logo reproduction with the source material.
Identity and anatomy
Check hands, small anatomical details, and identity consistency across multiple characters.
Reference-heavy edits
When using several reference images, verify which visual details were preserved in the result.
Pricing
GPT Image 2.5 Pricing
Choose a one-time PWZ credit pack for GPT Image 2.5 generation and editing. The exact credit cost is shown with your selected model and generation settings before you create.
600 credits · $0.042 per credit
Perfect for trying out PWZ — get started with a credit balance for your first generations.
- 600 credits
- High-resolution output support
2,000 credits · $0.030 per credit
For creators generating regularly — better per-credit value for daily PWZ workflows.
- 2,000 credits
- About 28% lower cost per credit
4,000 credits · $0.025 per credit
Best value for teams and content pipelines — more generations at the lowest per-credit cost.
- 4,000 credits
- About 40% lower cost per credit
- Priority processing speed
FAQ
GPT Image 2.5 FAQ
What is GPT Image 2.5?
GPT Image 2.5 is an OpenAI image model for generating and editing images from prompts and reference inputs. It includes Flare and Sunburst variants for different speed and precision needs.
What is the difference between GPT Image 2.5 Flare and Sunburst?
Flare is designed for faster generation, while Sunburst spends longer on requests where added precision and fine detail matter. Both support image generation and editing.
How do I use GPT Image 2.5?
Enter a prompt, optionally upload up to 10 reference images, choose Flare or Sunburst, select your output settings, and generate. For an edit, describe what should change and what should stay unchanged.
Can GPT Image 2.5 edit an existing image?
Yes. On PWZ, switch to Image to Image, upload up to 10 reference images, and describe the targeted changes you want.
Does GPT Image 2.5 support transparent backgrounds?
Yes. Choose Transparent under Background and select PNG or WebP to preserve transparency.
What image sizes does GPT Image 2.5 support?
On PWZ, GPT Image 2.5 supports Auto sizing plus 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, and 9:16 aspect ratios for square, landscape, and portrait images.
Is GPT Image 2.5 better than GPT Image 2?
GPT Image 2.5 adds Flare and Sunburst choices on PWZ, while the best model still depends on your prompt, references, quality setting, and preferred result.
Ready to create
Create Your Next Image With GPT Image 2.5
Start with a prompt, add a reference when you need more control, and choose Flare for faster iteration or Sunburst for added detail and precision.
Generate With GPT Image 2.5