zhineng.dev
English
Get started
← All models

GPT-Image-2.6 Sunburst

Published Publisher OpenAIgpt-image-2.6-sunburst

Overview

GPT-Image-2.6 Sunburst is OpenAI’s image generation and editing model for detailed compositions and precise visual work. It accepts text and image inputs and produces images. OpenAI positions Sunburst for greater detail and precision, while Flare is optimized for faster iteration. (developers.openai.com)

Capabilities and uses

  • Generate images from prompts and edit existing images, including edits guided by visual markup.
  • Produce animated GIFs and SVG graphics through the Image API. GIFs can be used for looping character or product animations; SVG output can be used for graphics such as icons and illustrations.
  • Adjust complexity to low, medium, or high independently of image quality. The default complexity is high; higher complexity can help with dense text and layouts but may increase latency or cause unwanted changes.
  • Suitable for campaign creative, product imagery, brand assets, and complex visual concepts. (developers.openai.com)

Limitations

  • Text is input-only; audio and video are not supported as model modalities. Animated GIF output is supported separately. (developers.openai.com)
  • Precise text placement, complex layouts, and consistency across generations can still be imperfect. Complex prompts may take up to two minutes; prompts and generated images are subject to content moderation. (developers.openai.com)
  • GIF output supports 2–16 frames and does not support streaming or partial images. SVG output is supported for generation, not the image-edit endpoint, and does not support streaming or partial images. (developers.openai.com)

API pricing

OpenAI lists the following USD per 1 million tokens for this model, checked November 14, 2026: (developers.openai.com)

Token type Input Cached input Output
Text $5.00 $1.25 Not applicable; text is input-only
Image $8.00 $2.00 $30.00

Cached-input rates apply to eligible cached usage; the applicable caching conditions for this model were not independently verified. Adjusting complexity has no separate fee. Markup-guided edits use the same token rates as other edits. SVG output also incurs charges for the authoring model’s tokens. Total image cost depends on token usage, including input and output. (developers.openai.com)

Release and official sources