GPT Image 2.6
by OpenAI
Flare for fast iteration. Sunburst for detailed product imagery and precise edits.
GPT Image 2.6 is available in Zalify Canvas in two versions: Flare for fast everyday generation, and Sunburst for higher-fidelity results and tighter control over edits. Choose Sunburst for final product images with references, or Flare to explore creative directions quickly.
Both versions support image generation, editing and reference images, with five quality tiers: low, medium, high, extra high and max. Medium is the default. Output resolution is a separate choice of 1K, 2K or 4K, and automatic aspect ratio lets the model choose the composition.
The prices below are estimates for 1K square output. Actual charges depend on the generation cost and your workspace markup, so they may differ from the estimate. Compared with GPT Image 2.5, low and medium quality cost more; high and above cost the same.
GPT Image 2.6 pricing on Zalify
GPT Image 2.6 offers Flare for everyday speed and Sunburst for higher-fidelity work. Both share the estimates below for one 1K square image, including Zalify’s standard 1.5× markup. Choose from five quality tiers; medium is the default.
| Quality / resolution | Price per image |
|---|---|
| Low quality | $0.0372 |
| Medium quality | $0.0654 |
| High quality | $0.0792 |
| Extra high quality | $0.1407 |
| Maximum quality | $0.3162 |
These are estimates, not fixed per-image charges. The app estimates 2K at 2× and 4K at 2.5× the 1K rate for the same quality. Final billing uses the actual cost reported for each generation, including any input cost, multiplied by your workspace markup and rounded up to $0.0001. Your chosen size and generation inputs can affect the final charge.
What a typical render costs
One image at Low quality
$0.0372
Ten images at Low quality
$0.372
One image at Medium quality
$0.0654
Ten images at Medium quality
$0.654
One image at High quality
$0.0792
Ten images at High quality
$0.792
One image at Extra high quality
$0.1407
Ten images at Extra high quality
$1.407
One image at Maximum quality
$0.3162
Ten images at Maximum quality
$3.162
Specifications
Provider
OpenAI
Resolution tiers
Low quality · Medium quality · High quality · Extra high quality · Maximum quality
Max output
4K
Aspect ratios
1:1 · 9:16 · 16:9 · 3:2 · 2:3 · auto
Image editing
Yes
Reference images
Yes
Where GPT Image 2.6 shines
Two versions for fast iteration or higher-fidelity final assets
Reference images and editing for product-focused creative
Five quality tiers, with resolution selected separately
Square, portrait, landscape and automatic aspect ratios
Best for
Frequently asked questions
For a 1K square image, GPT Image 2.6 is estimated from $0.0372 at low quality to $0.3162 at Maximum quality, including the standard 1.5× markup. Flare and Sunburst share these estimates. Larger output costs more; final billing uses the actual generation cost and your workspace markup. Charges are deducted from your usage balance in US dollars; there is no per-model subscription.
No. Every Zalify workspace with usage balance can generate with GPT Image 2.6 in Zalify Canvas. Paid plans include a monthly usage allowance; you can also add balance at any time, and purchased balance never expires.
Yes. The generate button in Zalify Canvas shows an estimate for the exact model, resolution and length you selected — for example "Generate video · est. $0.24" — before you commit.
No. You are only charged for renders the provider actually completes. If a generation errors out, nothing is deducted from your balance.
Every model in Zalify Canvas is priced on the same USD basis, so you can compare directly on the AI models overview. Cheaper tiers are ideal for iterating on a concept; switch to the model or resolution you want for the final render.
Other image models
GPT Image 2.5
From $0.009 per image
GPT Image 2
From $0.0165 per image
Seedream 4.5
From $0.06 per image
Seedream 5.0 Pro
From $0.0675 per image
Nano Banana Pro
From $0.201 per image
Try GPT Image 2.6 in Zalify Canvas
Start a 3-day free trial, add balance when you need it, and generate with every model from one canvas — no per-model subscriptions.