GPT Image 2.5 AI Image Generator & Photo Editor

Q: What is GPT Image 2.5 (Flare & Sunburst)?

Soutine provides access to OpenAI's GPT Image 2.5 models via fal.ai inference, featuring Flare (optimized for fast everyday text-to-image and reference photo editing) and Sunburst (tailored for precision typography and intricate detail).

Q: Does GPT Image 2.5 support native transparent backgrounds?

Yes. Select a transparent background to generate a PNG image with a native alpha channel.

Q: How does reference photo editing (image-to-image) work?

In edit mode, you can upload up to 4 reference images with natural-language instructions up to 1,000 characters. The model modifies the requested elements—such as attire, lighting, or specific objects—while preserving the core subject and composition without requiring manual mask painting.

Q: What resolutions and aspect ratios are supported in Soutine?

Soutine supports 5 fixed standard production aspect ratios: square 1:1 (1024×1024), landscape 4:3 (1024×768) and 16:9 (2560×1440), and portrait 3:4 (768×1024) and 9:16 (1440×2560), across low, medium, and high quality presets. Custom pixel inputs and auto-dimensions are not supported.

Q: Can I use generated images for commercial projects?

Commercial use is subject to OpenAI content policy, Soutine terms of service, your rights in any uploaded inputs, and applicable laws. Users may utilize their outputs for creative and commercial projects, but the platform cannot provide copyright guarantees.

Q: What is the difference between Flare and Sunburst in this workspace?

Flare delivers lower latency for rapid creative ideation, concept exploration, and responsive photo editing. Sunburst allocates more inference compute to fine details, rendering dense typography, intricate labels, and complex compositions.

Q: Which export formats and outputs are supported?

Each run generates one PNG image. Choose low, medium, or high quality; PNG also supports transparent backgrounds.

Generate and edit images using GPT Image 2.5 Flare and Sunburst. Create natural-light compositions, render transparent background PNG assets, or refine photos with multi-reference editing.

Text to Image

Upfront credit estimation before generation · browse gallery showcase →

Verified Edit Before & After

GPT Image 2.5 Targeted Image Editing

A targeted edit changes the sphere color while preserving the poster layout.

Input

Result

Coral to cobalt

Starting from the Flare poster, change the coral glass sphere to cobalt blue while keeping the typography and lighting.

Prompt

Change only the coral glass sphere to cobalt blue glass. Keep the headline, footer, paper sculpture, lighting, camera, composition, and every other detail unchanged.

Feature deep-dive

GPT Image 2.5 Before & After Showcase

Upload a reference photo and describe the change. GPT Image 2.5 keeps identity, pose, and scene details, and edits the part you name.

Identity

Formalwear, same child

The red sweater becomes a white tuxedo with a black bow tie. Face, pose, printed photo, and the room around it stay put.

Restyle

Pet photo to stunt-dog

A snapshot of a dog becomes an action-movie stuntman. Face, markings, and expression stay recognizable.

Scene

Same bedroom, made bed

The messy room is tidied and the bed is made. Furniture, lighting, and camera angle stay the same.

Showcase & Examples

GPT Image 2.5 Prompts & Showcase Gallery

Explore 15 creative GPT Image 2.5 prompt ideas for posters, product photography, typography labels, and reference editing. Choose a prompt to try in the generator.

Text to image

Retro Cat Magazine

Create a Japanese vintage magazine poster with film grain. An orange cat wearing headphones listens to music against a muted deep blue background. Large orange-red headline: "LUCAS". Small footer: "Seize The Moment". Leave space above the head. Premium retro print aesthetic, portrait composition.

Image to image

Photo to 3D Clay Character

Transform the subject in the uploaded photo into a stylized 3D clay character. Preserve the recognizable identity, composition and framing. Use soft clay materials, rounded forms, large expressive eyes, simplified anatomy and a playful pastel palette. Bold blue studio background, soft frontal lighting, subtle shadows, polished render.

Text to image

Paris Miniature Weather Card

Create a 45-degree isometric miniature 3D Paris scene with the Eiffel Tower, Notre-Dame, Seine and Parisian buildings. Refined PBR textures, soft winter daylight, light snow and a clean pale background. Center the title "Paris", a cloudy weather icon, "January 3, 2026" and "4°C" above the city. Square composition; illustrative sample weather, not a live forecast.

Text to image

Brand Identity Bento Grid

Create a professional bento-grid brand identity presentation for "BASE", a retro burger restaurant. Include a burger campaign image, social post mockup, logo construction grid, macro food poster, packaging mockup and four color swatches. Cohesive cream, red, teal and brown palette, bold typography, checkerboard accents, premium photorealistic mockups, landscape layout.

Text to image

Floating Sandwich Advertisement

Create a premium food advertisement of a sandwich labeled "Fresh Stack Club" floating in an exploded tilted view. Separate toasted bread, lettuce, tomato, cheese and grilled chicken into detailed layers, with realistic dripping sauce and flying crumbs. Bright sunlit kitchen background, shallow depth of field, fresh appetizing colors, wide composition.

Text to image

Floating Japan Travel Poster

Create a cinematic travel poster of a floating miniature island shaped like Japan above white clouds. Include Mount Fuji, Tokyo Tower, a red torii gate, traditional temples, cherry blossoms and forested mountains. Large white 3D letters read "JAPAN". Aerial perspective, vivid natural colors, warm sunlight and detailed miniature scenery, portrait composition.

Text to image

Watercolor Forest Stream

Paint a peaceful forest clearing with a stream winding through grass and wildflowers. Delicate translucent watercolor, wet-on-wet washes, finely detailed irregular foliage, clear organic edges without ink outlines. Sunbeams filter through the canopy with soft volumetric daylight. Fresh greens, luminous water, airy composition, landscape format.

Image to image

Photo to Watercolor Portrait

Use the uploaded photo as the facial reference, preserving recognizable features and identity. Turn it into a detailed ink and watercolor portrait on white paper, with expressive blue, orange and purple splashes, pigment blooms and fine drips. A realistic hand with a paintbrush enters the foreground, as if painting the subject’s sleeve. Soft lighting and a clear hand-painted finish.

Text to image

Citrus Skincare Campaign

Photorealistic skincare advertisement: a frosted yellow glass serum bottle labeled "Glow Recipe – Lemon Brightening Glow Serum (30 ml)" floats above fresh lemon halves. A clear water splash curves behind it. Soft lavender studio background, luminous glass, condensation and crisp citrus textures, elegant portrait composition.

Text to image

AI FUTURE Perspective Type

Design a dramatic typographic poster reading "AI FUTURE" in heavy condensed sans-serif lettering. Acid yellow letters on matte charcoal gray. Duplicate the letters along receding perspectives toward a vanishing point, with repeated thin white outlines creating depth and motion. Bold graphic energy, portrait layout, no additional text.

Image to image

Clothing to Catalog Photo

Extract the outfit from the uploaded photo and create a clean e-commerce product photograph. Remove the person completely while preserving the garment design, colors, seams, folds and fabric texture. Keep a natural three-dimensional garment shape as on an invisible mannequin. Center it on a pure white background with professional soft studio lighting.

Image to image

Photo to Layered Papercraft

Turn the uploaded photo into a handcrafted layered paper-cut illustration. Preserve its subjects and composition. Use rounded shapes, cute simplified faces, matte cardstock textures and visible stacked layers with soft shadows. Give each main character a thick white cut-paper border. Muted blue, green and warm neutral colors, cozy storybook mood, soft even lighting.

Text to image

Recipe Infographic

Design a clean recipe infographic for chicken avocado wraps. Center two appetizing cut wraps floating in perspective. Around them place illustrated ingredients with quantities: 2 tortillas, 1 cooked chicken breast, 1 avocado, 1 tomato, lettuce and yogurt sauce. Show three numbered steps: prepare, assemble, roll. Clear headings, small food icons, fresh natural colors, soft shadows, airy white background and readable typography.

Image to image

Restore and Colorize a Photo

Restore and colorize the uploaded old photograph with natural modern photographic clarity. Remove scratches, creases, stains, noise and fading. Preserve the exact facial features, identities, expressions, clothing, pose, background and original framing. Recover realistic skin tones and fine fabric detail. Return only the restored photograph, with no comparison layout, invented objects or plastic-looking skin.

Text to image

New York Four Seasons

Create a colorful cut-paper collage infographic for a third-grade classroom about New York City seasons. Title: "NYC SEASONS". Four panels labeled "SPRING", "SUMMER", "AUTUMN", "WINTER" show blossoms, sunny parks, autumn leaves and snowy streets. Include recognizable city landmarks, cheerful children and one short simple explanation per season. Playful handmade textures, clear readable lettering, landscape composition.

Core Capabilities

What Can GPT Image 2.5 Create?

Flare Fast Inference

Low-latency generation pipeline optimized for rapid creative ideation and everyday image editing.

Native Transparent Backgrounds

Generate isolated cutouts and product mockups with native Alpha channels in PNG format.

Reference Photo Editing

Upload up to 4 reference images to guide stylistic adjustments, wardrobe changes, or background modifications.

Model Selection

GPT Image 2.5 Flare vs Sunburst Benchmark

Both models share Soutine workspace features—text and reference photo editing, 5 fixed standard aspect ratios, and transparent PNG exports—with distinct optimizations for turnaround speed versus dense detail.

Test in generator above

Workflow & Capability	GPT Image 2.5 Flare	GPT Image 2.5 Sunburst
Primary Creative Focus	✓	High-precision typography, micro-labels & intricate art
Turnaround & Generation Latency	✓	Extended generation time for dense detail passes
Native Transparent Background (PNG)	✓	Supported in PNG
Reference Photo Editing (Up to 4 images)	✓	Supported with high detail adherence
Standard Aspect Ratios (1:1, 4:3, 3:4, 16:9, 9:16)	✓	Same 5 fixed standard aspect ratios
Upfront Credit Estimation & Pricing	✓	Credits are estimated before generation based on the mode, quality, aspect ratio, and number of reference images, with applicable membership discounts. Each run generates one image.

Workflow Guide

How to Generate & Edit with GPT Image 2.5

Follow five clear steps to move from idea to production-ready graphic.

Select text-to-image or upload reference

Start from a blank canvas with a descriptive prompt up to 1,000 characters, or upload up to 4 reference images to enter edit mode.

Specify background and quality presets

Choose opaque or transparent background, and set quality (low, medium, or high) to balance turnaround and detail level.

Choose standard aspect ratio

Choose 1:1, 4:3, 3:4, 16:9, or 9:16. Each run generates one image.

Submit task via asynchronous queue

Your request is dispatched through the asynchronous fal queue with real-time status tracking.

Download your PNG

Download the generated PNG, or upload it again for another edit.

FAQ

GPT Image 2.5 Frequently Asked Questions

Verified details regarding GPT Image 2.5 capabilities, pricing, and output formats in Soutine.

What is GPT Image 2.5 (Flare & Sunburst)?

How does pricing and workspace credits work?

Does GPT Image 2.5 support native transparent backgrounds?

How does reference photo editing (image-to-image) work?

What resolutions and aspect ratios are supported in Soutine?

Can I use generated images for commercial projects?

What is the difference between Flare and Sunburst in this workspace?

Which export formats and outputs are supported?

Get Started

Ready to Create with GPT Image 2.5?

Experience fast Flare inference, precision Sunburst typography, and reference editing directly in the workspace.

Open Generator Above

Upfront credit estimation before generation 5 fixed production aspect ratios Shared workspace history