OpenAI
Newreleased September 2026GPT Image 2.5
OpenAI's newest image model
GPT Image 2.5 follows complex instructions, renders dense legible text and edits with multiple reference images. Choose from low to max quality depending on the job.
- Five quality levels
- Edits with multiple reference images
- Dense, legible text
- Up to 4 images per request
- Min. credits
- 3
- Typical time
- ~25s
Specs
Capabilities
- Modes
- Text to imageImage to image
- Inputs
- TextPhoto
Up to 10 reference images
- Aspect ratios
- 16:99:161:14:33:43:22:3
Credits
Pricing
Credits per output. Free generations cover Standard quality.
| Setting | Quality | Credits |
|---|---|---|
| low quality | Standard | 3 |
| medium quality | Standard | 10 |
| high quality | High | 30 |
| xhigh quality | Ultra | 50 |
| max quality | Ultra | 100 |
![FLUX.2 [pro]](https://api.osonai.uz/storage/demo/showcase/images/flux-2-pro-04.jpg.w1080.webp)