Drop files here, paste files, or add a URL. JPEG, PNG or WebP, up to 10 MB per image and 16 images.
Drop files here, paste files, or add a URL. Use an HTTP(S) mask URL together with at least one reference image.
Example Output
This is sample data. Generate a new result to see real output.
Transparent pricing with no hidden fees. Pay as you go.
GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5GPT Image 2.5| Model | Spec | PoYo price | Official price | You save |
|---|---|---|---|---|
| GPT Image 2.5 | low quality - 1K Flare / Sunburst | $0.0066/gen 1.31 credits /gen | - | - |
low quality - 2K Flare / Sunburst | $0.013/gen 2.65 credits /gen | - | - | |
low quality - 4K Flare / Sunburst | $0.012/gen 2.47 credits /gen | - | - | |
medium quality - 1K Flare / Sunburst | $0.015/gen 2.95 credits /gen | - | - | |
medium quality - 2K Flare / Sunburst | $0.030/gen 5.99 credits /gen | - | - | |
medium quality - 4K Flare / Sunburst | $0.029/gen 5.81 credits /gen | - | - | |
high quality - 1K Flare / Sunburst | $0.059/gen 11.78 credits /gen | - | - | |
high quality - 2K Flare / Sunburst | $0.120/gen 23.94 credits /gen | - | - | |
high quality - 4K Flare / Sunburst | $0.112/gen 22.39 credits /gen | - | - | |
xhigh quality - 1K Flare / Sunburst | $0.105/gen 20.93 credits /gen | - | - | |
xhigh quality - 2K Flare / Sunburst | $0.213/gen 42.55 credits /gen | - | - | |
xhigh quality - 4K Flare / Sunburst | $0.199/gen 39.78 credits /gen | - | - | |
max quality - 1K Flare / Sunburst | $0.236/gen 47.11 credits /gen | - | - | |
max quality - 2K Flare / Sunburst | $0.479/gen 95.74 credits /gen | - | - | |
max quality - 4K Flare / Sunburst | $0.448/gen 89.51 credits /gen | - | - |
* Actual fees are based on the final output.
High-fidelity image generation, precision editing, and consistent creative revisions
GPT Image 2.5 turns text prompts and reference images into product photos, campaign visuals, portraits, and design assets. Its image-to-image editing preserves recognizable subjects, refines selected elements, and maintains detail across successive revisions.
Flare focuses on fast text-to-image generation and everyday creative iteration. Sunburst focuses on precise image editing and detailed production assets. Explore both versions and compare their capabilities with GPT Image 2 on PoYo.
Flare is OpenAI's default choice for everyday image applications. It accepts text and image inputs and supports image generation, editing, and inpainting. Its focus is fast iteration for social content, product experiences, image prototypes, and high-volume creative workflows.
Sunburst is OpenAI's most capable image generation and editing model. It accepts text and image inputs and supports inpainting. Choose it for detailed campaign assets, polished product imagery, and editing workflows that prioritize precision over generation speed.
Text-to-image generation, reference image fidelity, inpainting, and multi-turn editing for product imagery and creative assets.
01
GPT Image 2.5 API uses reference images and text instructions to create new scenes while preserving recognizable people, products, and visual identities. High-fidelity image-to-image generation retains defining facial traits and product details as the background, lighting, or artistic style changes. Build portrait variations, e-commerce product photos, and branded creative assets around the same source image.
02
GPT Image 2.5 API supports precise image editing and inpainting for changes to selected parts of an existing image. Provide the source image and describe the product, background element, or copy to update. The model follows the localized edit while preserving surrounding composition and visual treatment, supporting product retouching, campaign revisions, and targeted background changes.
03
GPT Image 2.5 API improves multi-turn image editing by maintaining earlier changes and image detail through successive revisions. Start with a generated image or reference photo, then adjust the background, objects, and composition in separate steps. Using each result as the next reference keeps creative feedback specific and supports subject consistency across the full editing workflow.
04
GPT Image 2.5 API follows detailed visual instructions for structured layouts, artistic styles, and transparent backgrounds. Specify the subject placement, visual hierarchy, requested text, and background treatment in the prompt. Use text-to-image generation to create presentation graphics, poster concepts, and compositing assets, then apply image editing to refine individual elements within the chosen layout.
05
GPT Image 2.5 Flare API focuses on fast text-to-image generation for everyday creative workflows. OpenAI reports 50% lower generation latency than GPT Image 2 alongside image-quality improvements. Use Flare to explore campaign concepts, social graphics, and product variations in repeated generation cycles. Sunburst serves precision-focused editing and detailed creative production with longer generation times.
Use GPT Image 2.5 with product reference photos to create new backgrounds, seasonal compositions, and listing visuals. Targeted edits let teams revise a scene while retaining the product's defining details.
Develop campaign concepts with Flare, then use Sunburst for detailed image revisions. GPT Image 2.5 supports variations in composition, background, and visual treatment around a consistent creative brief.
Create reference-led portraits, promotional graphics, and themed posts with GPT Image 2.5. Use Flare to explore ideas quickly and make focused changes to the selected image without restarting the composition.
Translate layout instructions, artistic direction, and reference assets into image concepts with GPT Image 2.5. Successive edits support review cycles that adjust individual elements while retaining earlier design decisions.
Create illustrations, presentation graphics, and teaching visuals from a structured brief. GPT Image 2.5 supports specific compositions and styles; editors can revise individual elements and review factual details and text before publication.
Design image workflows around text prompts, reference inputs, generation, and localized editing. GPT Image 2.5 provides Flare for everyday iteration and Sunburst for applications where detailed editing is the primary requirement.
GPT Image 2 and GPT Image 2.5 both generate and edit images from text and image inputs. GPT Image 2.5 builds on that workflow with improved reference fidelity, more precise editing, stronger consistency across revisions, and faster generation through Flare.
| Capability | GPT Image 2.5 | GPT Image 2 |
|---|---|---|
| Inputs and output | Text and image inputs; image output through Flare or Sunburst | Text and image inputs; image output |
| Reference image fidelity | Improved preservation of recognizable subjects, lighting, and textures | High-fidelity image inputs for reference-led generation and editing |
| Local editing | More precise targeted edits with improved preservation of surrounding details | Image editing and inpainting for existing images |
| Successive revisions | Improved retention of earlier edits and image quality across multiple turns | Image editing using an existing image as input for each revision |
| Layouts and style | Improved complex visual instruction following, layouts, and transparent backgrounds | Prompt-guided image generation and editing with flexible image sizes |
| Generation speed | Flare: 50% lower latency than GPT Image 2, as reported by OpenAI; Sunburst prioritizes precision with longer generation times | Baseline for OpenAI’s published Flare latency comparison |
Comparison based on OpenAI’s GPT Image 2.5 announcement and GPT Image 2 model documentation. The 50% latency reduction refers specifically to Flare compared with GPT Image 2.
Step 1: Prepare the prompt and reference images
Describe the subject, composition, lighting, style, and required text. For edits, provide the source image and name the elements to change or preserve.
Step 2: Select Flare or Sunburst
Choose Flare for everyday generation and rapid iteration. Choose Sunburst when detailed image editing is the priority.
Step 3: Generate, review, and refine
Inspect the output for subject fidelity, layout, and text. Use the previous result as the reference for each targeted revision.
PoYo brings image model listings and API developer resources together in one place. Review GPT Image 2.5 capabilities, input options, and use cases on the same platform to prepare your image application’s integration plan.
Compare the generation speed and editing focus of the two GPT Image 2.5 variants on PoYo. Choose the version that fits your creative requirements for product images, campaign assets, or reference image editing.
PoYo provides API documentation and request examples covering authentication, task submission, and result handling, helping developers prepare the integration logic for image applications.
PoYo’s asynchronous task system supports status polling and Webhook callbacks. Applications can handle request submission, task waiting, and result retrieval separately to organize image generation and editing workflows.
