GPT Image (OpenAI) API
OpenAI's GPT Image 2 — photorealistic generation with sharp text rendering
Overview
GPT Image 2 is OpenAI's next-generation image model built for photorealism, clean editing, and sharp text rendering. Access through AI Generate with pay-per-image billing.
All 10 GPT Image (OpenAI) variants & pricing
Every variant is accessible via the same API endpoint. Switch variants by changing the model parameter. No extra setup.
| Variant | Credits | Price USD | Unit | |
|---|---|---|---|---|
| gpt image 1.5, image-to-image, medium | 4.00 | $0.0280 | per image | Details → |
| gpt image 1.5, text-to-image, medium | 4.00 | $0.0280 | per image | Details → |
| gpt image 2, image-to-image, 1k | 6.00 | $0.0420 | per image | Details → |
| gpt image 2, text-to-image, 1k | 6.00 | $0.0420 | per image | Details → |
| gpt image 2, image-to-image, 2k | 10.00 | $0.0700 | per image | Details → |
| gpt image 2, text-to-image, 2k | 10.00 | $0.0700 | per image | Details → |
| gpt image 2, text-to-image, 4k | 16.00 | $0.1120 | per image | Details → |
| gpt image 2, image-to-image, 4k | 16.00 | $0.1120 | per image | Details → |
| gpt image 1.5, image-to-image, high | 22.00 | $0.1540 | per image | Details → |
| gpt image 1.5, text-to-image, high | 22.00 | $0.1540 | per image | Details → |
Use GPT Image (OpenAI) via API
Example call with the default variant. Replace model with any slug from the table above.
curl https://aimarcusimage.eu/api/v1/jobs/createTask \
-H "Authorization: Bearer sk-aig-..." \
-H "Content-Type: application/json" \
-d '{
"model": "openai-gpt-image-1-5-image-to-image-medium",
"input": {
"prompt": "a red fox sitting in snow, golden hour, photographic"
}
}'
FAQ
How much does GPT Image (OpenAI) cost?
Pricing starts at $0.0280 per image for the most affordable variant. You pay per call from your credit balance — no monthly subscription, no minimum commitment. Starter top-up is $10 → 1,430 credits.
Is this an official OpenAI API?
AI Generate is an AI model aggregator that routes calls to OpenAI's upstream API. You get the exact same model outputs with simpler billing (pay per call, no contracts), one API key for 307 models, and the same OpenAI-compatible request shape across all providers.
Can I use GPT Image (OpenAI) in production?
Yes. The endpoint is rate-limited to 20 requests per 10 seconds per API key, retries are idempotent via taskId, and async results are delivered via polling or webhook. Set a spend cap in dashboard settings to protect against runaway usage.
Which GPT Image (OpenAI) variant should I pick?
For first tests, start with the cheapest variant (lowest "Price USD" in the table above). Upgrade to higher-resolution or "Quality" tiers only when you have validated prompts. The Playground lets you A/B test all variants with the same per-call pricing as the API.