Skip to main content

GPT Image 2.6 Flare

Generate a picture, then make a deliberate edit.

Try
Heirloom plaid Christmas
Heirloom plaid Christmas

A candid snapshot from a late-2000s compact digital camera with the flash firing straight on, taken in an old family house on Christmas evening 2026. Vertical portrait, waist-up. The main subject is an original woman in her mid-twenties with long straight auburn hair and light skin, sitting on the arm of a worn leather sofa, facing the camera with a relaxed smile, face clearly visible, wearing a cream cable-knit sweater with a muted green and navy tartan blanket over her lap, holding a chipped enamel mug. A wood-panelled living room behind her with a lit stone fireplace, an evergreen garland on the mantel, aged brass candlesticks and an old brass lamp giving warm tungsten light. A friend in a grey sweater sits half out of frame on one side, and an older relative is a soft blur by the fireplace. Mixed light: warm orange tungsten in the background, cold flat flash on the faces. Film-like grain and noise, slight motion blur on the people behind, crooked framing. Cosy, lived-in, slightly cluttered. Real people, natural skin texture, not retouched. No text, no logos, no brands, no famous people, no Santa, no clip-art decorations.

GPT Image 2.6 Flare

How it works

Try it now
  1. 01

    Say what you want

    A sentence or two is enough.

  2. 02

    Use GPT Image 2.6 Flare

    Choose the settings that fit what you want to make.

  3. Heirloom plaid Christmas
    03

    Make it

    Start creating, then keep working with HowWorks.

Explore the possibilities

Heirloom plaid Christmas
GPT Image 2.6 Flare
After: A candlelit Christmas windowBefore: A candlelit Christmas window
GPT Image 2.6 Flare

GPT Image 2.6 Flare at a glance

Good at
Fast GPT Image generation, transparency, precise edits and SVGs.
Modes
Text to Image, Image Editor
Resolution
1K, 2K, 4K
Aspect ratios
Auto, 1:1, 4:3, 3:4, 16:9, 9:16
Reference images
Up to 16
Output images per request
1–10
Quality
Low, Medium, High, Xhigh, Max, Auto

About

GPT Image 2.6 Flare brings image generation and editing into the same project workflow. Start from a written brief, or attach an image and explain what should change. A candlelit shop window, a different background or a revised composition is easier to judge when you also say what must remain: the product shape, palette or relative position of objects. Current controls include the supported dimensions, quality and output options. Compare the result with the source, especially for text, fine geometry and anything that carries your product identity.

Questions

How do I use GPT Image 2.6 Flare in HowWorks?

Start with the model selected on this page. Write the brief, add the inputs required by your chosen operation and review the current settings before sending.

Which inputs and settings are supported?

The model specifications and controls on this page show the supported inputs and output settings. Availability can differ by operation; selecting a model does not enable every capability advertised elsewhere.

Does selecting this model change every AI tool in my project?

No. It selects the image generator for this task. The Agent conversation model and other generation tools have their own choices.

How should I evaluate the result?

Compare it with your brief and any source material. Check composition, fine detail, lettering and the intended output format.