Travel Poster
3:2A vintage travel poster for Kyoto featuring cherry blossoms over a temple, Japanese text reading '京都へようこそ', retro mid-century illustration style, gold and pink palette
GPT Image 1.5
ID:
openai/gpt-image-1.5GPT Image 1.5 image generation API by OpenAI — 4x faster than v1 with superior text rendering, 3 aspect ratios, and medium/high quality modes.
Text to imageImage editingReference imagesText rendering
Input
OutputView all
Output will appear here
medium
$0.03
high
$0.16
cURL
curl https://api.runbase.net/v1/runs \
-H "Authorization: Bearer $RUNBASE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "openai/gpt-image-1.5",
"input": {
"prompt": "A cinematic product photo of a ceramic lamp",
"aspect_ratio": "1:1",
"resolution": "1K"
}
}'Examples
Product Photography
3:2A photorealistic product shot of a luxury watch on a marble surface, studio lighting with soft reflections, shallow depth of field, magazine advertisement quality
Children's Book Illustration
3:2A children's book illustration of a fox and owl reading together under a giant mushroom, watercolor style, warm autumn colors, whimsical and cozy atmosphere
Overview
GPT Image 1.5 is an incremental update released in December 2025 that prioritizes speed and text accuracy. It generates images roughly 4x faster than GPT Image 1 at lower cost, making it practical for high-throughput production work. Text rendering is significantly improved — especially dense, small, or multilingual text — and editing precision is better at preserving logos and faces in reference images.
Use cases
Poster and infographic design with dense embedded text. UI mockups and wireframe visualizations. High-throughput production workflows where turnaround time matters. Image editing that needs to preserve existing logos, faces, or brand elements in reference images.
Inputs
All parameters are passed in the input object of the run request.
| Parameter | Required | Description |
|---|---|---|
| prompt | Yes | Text description of the image to generate (1–20000 chars) |
| aspect_ratio | No | Output aspect ratio. Default 1:1. Options: 1:1, 3:2, 2:3 |
| quality | No | Output quality. Default medium. Options: medium, high |
| image_urls | No | Up to 4 reference images (max 4 MB each) for image editing |
Prompt tips
Use the full prompt length for text-heavy designs
With a 20,000-character prompt limit, you can spell out exact text strings, font descriptions, and layout instructions. This is especially useful for poster or infographic prompts with multiple text blocks.
Choose quality mode deliberately
medium is fast and cost-effective for iteration. Switch to high for final deliverables where fidelity matters — the visual difference is most noticeable in fine detail and text sharpness.
Limitations
- No resolution control — quality mode affects fidelity, not pixel dimensions
- 3 aspect ratios (fewer options than GPT Image 2)
- Maximum 4 reference images per request, 4 MB each
- Complex layouts with many overlapping elements may still need manual review
- Prompt length capped at 20,000 characters
FAQ
How is GPT Image 1.5 different from GPT Image 1?
It is roughly 4x faster, more cost-effective at medium quality, and has significantly better text rendering. It also adds a quality toggle (medium/high). Both support the same 3 aspect ratios (1:1, 3:2, 2:3).
Does it support resolution control?
No. Unlike GPT Image 2, there is no explicit resolution parameter. Use the quality setting (medium or high) to control output fidelity.
Can I edit existing images?
Yes. Upload up to 4 reference images via image_urls. The model is particularly good at preserving logos and faces during edits.