Comparison
GPT Image 2.5 on Black Friday: 10,000 2K Images at 3 Cents Each
Black Friday math per image, plus per-image prices on 7 platforms
On this page
GPT Image 2.5 on HiAPI is priced per image, one price per resolution: $0.025 at 1K, $0.04 at 2K, $0.06 at 4K. Flare and Sunburst cost the same, and so do generation and editing. During Black Friday week (Nov 24 to Nov 30, PT), 10,000 2K images on HiAPI end up costing about $300, or 3 cents each.
By volume: 100, 2,000 and 10,000 images
| What you make | List price | How Black Friday applies | What it really costs | Per image |
|---|---|---|---|---|
| 100 images at 1K | $2.50 | Too small for the bonus or spend-back | $2.50 | $0.025 |
| 2,000 images at 2K | $80 | Top up $100, get $8; spend $80, get 5% ($4) back | About $68 | About $0.034 |
| 10,000 images at 2K | $400 | Top up $400 at once, get $40; spend $400, get 15% ($60) back | About $300 | $0.030 |
"What it really costs" is what you paid in, minus what's left in your balance afterwards (unused bonus plus the spend-back that lands on Dec 3). Image, video and audio spend count together toward the spend-back tier, so a week with video jobs too gets a higher tier.
Price per image on 7 platforms
Public list prices we checked on each platform on the evening of Nov 10 (Beijing time). Where a platform has quality tiers, we used the highest:
| Platform | 1K | 2K | 4K |
|---|---|---|---|
| APIMart (ext route) | $0.0085 | $0.03 | $0.04 |
| HiAPI | $0.025 | $0.04 | $0.06 |
| KIE | $0.03 | $0.05 | $0.08 |
| Pollo | $0.0528 | $0.1068 | not listed |
| WaveSpeed | $0.09 | $0.15 | $0.27 |
| Flaq | $0.10 | $0.16 | $0.28 |
| Replicate | $0.128 (any resolution) | same | same |
To be straight about it: APIMart's ext route is cheaper per image than HiAPI. At the $500 Black Friday tier HiAPI works out to about $0.019 at 1K and $0.031 at 2K, still above APIMart at 1K and about even at 2K. HiAPI is cheaper than KIE, Pollo, WaveSpeed, Flaq and Replicate.
Two more notes:
- fal charges per token ($30 per million image output tokens) and lists no per-image price, so it's not in the table.
- None of these platforms had announced a Black Friday deal as of Nov 10. The "X% below official" labels on their pages are standing prices.
Need control over quality? Use the @pro route
The per-image prices above don't vary by quality. If you need to set quality, HiAPI also offers gpt-image-2.5-sunburst@pro, priced by tier: low $0.0172, medium $0.0672, high $0.1829, xhigh $0.3572, max $0.7143. Run bulk work on the default route and save high and above for a handful of hero images.
Three ways to save on images this Black Friday
- Pull December's image work into this week. Product shot refreshes, ad variants, datasets: run them in one go.
- Top up in one payment. The bonus is per payment: $300 at once gets $30, three $100 top-ups get $24.
- Edits cost the same as generations. If you have a base image, go straight to image-to-image.
Model pages: GPT Image 2.5 Flare · GPT Image 2.5 Sunburst. Everyday pricing: GPT Image 2.5 API pricing. Budget math: How much to top up.
FAQ
Do Flare and Sunburst cost the same? Yes. They are two generation modes of GPT Image 2.5; pick the one whose style you need.
Does the per-image price drop on Black Friday? No. Black Friday adds a top-up bonus and spend-back; the per-image price stays the same, so the effective price drops.
When does the spend-back arrive? On Dec 3, in your balance, usable on every model.
Related articles
View all articles →-
Comparison
Gemini 4 Series and 3.8 Flash Model Selection: API, Cost, and Migration
Based on the Chat Completions-compatible API, compare the task positioning, reasoning effort, and token cost of the three model lines
-
Comparison
How to Choose gemini-4-carbon: A Selection Guide by Task, Context, Reasoning Effort, and Cost
Make decisions across four dimensions: task requirements, context, reasoning effort, and cost
-
Comparison
gemini-4-argon Model Selection and Migration Guide: API Capabilities, Context, and Cost
From API capabilities, context, and reasoning effort to cost, an actionable model selection framework for the same family