Models / OpenAI

GPT Image 2.6 Flare

Fast image generation and editing

Try GPT Image 2.6 Flare

OpenAI in Breme

One model, within your workflow

Use GPT Image 2.6 Flare in Canvas for image generation, inpainting, image expand, image editing, vector generation. Keep your prompts, references, and generated work together in your project, then take the result into the next stage of your production.

Capabilities and settings

Explore the inputs and controls available for each mode in Breme. Some options depend on the inputs and settings you choose; Canvas shows the combinations available for your request.

Image generation

Fast image generation and editing

Visual references
Up to 10 images
Prompt
Up to 32,000 characters
Aspect ratio
1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16
Quality
Low, Medium, High, Xhigh, Max, Auto
Background
Auto, Opaque, Transparent
File format
PNG, WebP, JPEG
Complexity
Low, Medium, High
Input requirements

These checks apply when their corresponding inputs or settings are used.

  • Image references must be recognized image files.
  • SVG files cannot be used as image references, so export a PNG or JPEG first
  • Transparent backgrounds require PNG or WebP output.

Inpainting

Fast image generation and editing

Source
One image, required
Selection mask
One image, required
Visual references
Up to 9 images or vectors
Prompt
Up to 31,750 characters
Complexity
Low, Medium, High

Image expand

Fast image generation and editing

Source
One image, required
Visual references
Up to 9 images or vectors
Prompt
Up to 31,750 characters
Complexity
Low, Medium, High

Image editing

Fast image generation and editing

Source
One image, required
Visual references
Up to 9 images
Prompt
Up to 32,000 characters
Aspect ratio
1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16
Quality
Low, Medium, High, Xhigh, Max, Auto
Background
Auto, Opaque, Transparent
File format
PNG, WebP, JPEG
Complexity
Low, Medium, High
Input requirements

These checks apply when their corresponding inputs or settings are used.

  • Image references must be recognized image files.
  • SVG files cannot be used as image references, so export a PNG or JPEG first
  • Transparent backgrounds require PNG or WebP output.

Vector generation

Editable SVG from a text prompt

Prompt
Up to 32,000 characters
Aspect ratio
1:1, 3:2, 2:3
Quality
Low, Medium, High, Xhigh, Max, Auto
Complexity
Low, Medium, High

From a prompt to your next piece

  1. Open Canvas and choose GPT Image 2.6 Flare from the model picker.
  2. Describe what you want to make. Add compatible references when the selected mode supports them, then choose your settings.
  3. Review the credit estimate before generating. Compare the result with your brief and refine your next request.

Before you generate

Review generated work before using it in a finished production. Results can vary between requests, and references guide the output without guaranteeing an exact match.

Model access and credit estimates depend on your plan and request settings. Check the options in your workspace before starting a generation.