GPT Image 2.6 Sunburst

GPT Image 2.6 Sunburst: Model Introduction & Practical Guide

GPT Image 2.6 Sunburst is OpenAI's image generation and editing model for detailed visual work. Build considered compositions, refine reference images, and prepare product, brand, and editorial concepts.

Quick Facts

Table of Contents

1. Model Overview
2. Core Features
3. Technical Specifications
4. Capability Comparison
5. Core Advantages
6. Recommended Use Cases
7. Example Prompts
8. Selection Recommendations
9. Pricing & API Notes

1. Model Overview

Sunburst emphasizes detailed compositions and image refinement. It is suited to work where relationships between objects, layout, and consistency across edits need careful review.

The model creates images from written instructions and uses reference images for directed revisions. Describe the intended subject, visual hierarchy, lighting, background, and required lettering. For an edit, identify the parts to preserve and the specific change to make.

OpenAI introduced the GPT Image 2.6 family with improvements to generation speed and consistency across edits. The two variants give teams different starting points for a visual workflow while keeping the same standard token rates.

2. Core Features

• Text-to-image creation: describe a visual concept and develop it into an image.
• Reference-guided editing: supply images to communicate a subject, composition, or desired revision.
• Quality control: choose an explicit quality level for a configured estimate, or Auto for usage-dependent quality.
• Canvas control: combine a resolution preset with an aspect ratio appropriate for the intended placement.
• Iterative refinement: inspect a result, then give the next request focused instructions and a suitable reference.

OpenAI also documents complexity control, markup-guided edits, SVG output, and animated GIF output in its Image API. The Studio controls available for this Item are the prompt, reference images, quality, resolution, aspect ratio, and batch size. The API feature list describes the model family; the visible panel defines the options available in this workflow.

3. Technical Specifications

A resolution preset and ratio select the image dimensions for a request. Choose the shape that fits the planned asset, then review the composition at its intended display size. Larger or more demanding outputs can use more tokens.

4. Capability Comparison

Choose by the result your project needs. Per-token pricing alone does not predict how many revisions or tokens a particular task will use.

5. Core Advantages

Use Sunburst when an image needs a more considered composition or a careful revision of an existing visual. Give the model a clear description of the intended result and the elements to preserve. Review subject identity, geometry, lettering, and fine details after each edit.

For any generated image, review small text, fine geometry, and the details that matter to the intended use. A reference communicates visual intent, but consistency should be checked in the returned image.

6. Recommended Use Cases

• Product and brand concepts where material, lighting, and composition need close review.
• Reference-guided refinements of a selected creative direction.
• Editorial illustrations with several subjects or a defined visual hierarchy.
• Campaign visuals that benefit from explicit instructions about preserved elements.

7. Example Prompts

New image

Create a landscape editorial illustration of a small design studio at dusk. Show three people reviewing a physical model around one table. Use warm task lighting, balanced spacing, and a restrained terracotta and navy palette.

Reference-guided edit

Using the reference image, preserve the product proportions, label placement, and camera angle. Change only the tabletop to dark walnut and soften the background lighting. Review the product lettering for accuracy.

These are starting prompts. Replace the subject, materials, framing, and output purpose with the requirements of your project.

8. Selection Recommendations

Choose Sunburst for detailed visual work and careful refinement. Choose Flare when interactive exploration is the main priority. Start with a manageable canvas and increase quality after selecting a composition. Shared token rates do not imply identical total usage per request.

Begin with the default medium-quality 1K configuration for a draft. Change one important variable at a time when comparing results. Use an explicit quality level when a pre-generation estimate is useful; Auto quality is dynamic and may have no fixed estimate before execution.

9. Pricing & API Notes

The standard text-input, image-input, and image-output rates used by this Item match OpenAI's published rates:

The Studio price panel and current pricing contract determine the charge for the selected configuration. The retained input policy counts cached input in the corresponding standard input category rather than applying a separate cache rebate.

An estimate uses configured usage budgets for the selected quality and resolution. Those budgets are planning assumptions, not measured tokens, a fixed fee per image, or a final price cap. Reference images contribute to usage. Final pricing uses the request's billable usage and the accepted pricing contract.

The underlying OpenAI model exposes image-generation and image-edit endpoints. Use the controls in Studio to create a new image or submit a reference-guided request.

Sources & Further Reading

• OpenAI model documentation
• OpenAI Images 2.6 announcement
• OpenAI image-generation guide
• OpenAI standard pricing