GPT Image 1.5 — AI Image Generator Online
OpenAI's previous-generation image model. Still available here until its API shutdown on December 1, 2026; after that this page switches to GPT Image 2.
GPT Image 1.5, released in December 2025, was OpenAI's flagship image model before GPT Image 2. Built into the GPT-5 architecture, it generates images 4× faster than GPT Image 1 at 20% lower cost, preserves original details during edits instead of re-rendering, and handles dense, small text accurately. OpenAI has deprecated it and will shut it down on December 1, 2026.
OpenAI has deprecated gpt-image-1.5 and will shut the API down on December 1, 2026. Until then it generates and edits normally here. From that day the generator on this page switches to GPT Image 2 automatically; your credits and history carry over. GPT Image 1 was already retired on October 23, 2026.
- 4× faster generation speed
- 20% cheaper than GPT Image 1
- Preserves original image elements during edits (no re-renders)
- Significantly better text rendering — dense, small, accurate
- Maintains visual consistency across multiple generations
- Handles complex scenes with multiple objects reliably
- Built directly into GPT-5 architecture
GPT Image 1.5 is 20% cheaper than GPT Image 1. Text input: $5/1M tokens. Image input: $8/1M tokens. Image output: $32/1M tokens. Typical cost: ~$0.01–$0.05 per image.
Production-grade use cases
Photoreal lead images for landing pages, decks, and editorials — precise composition, consistent lighting, magazine-quality output.
Generate sets of images that hold the same palette, lighting, and styling across multiple shots — ideal for content systems.
Spin up coordinated ad creatives in one session: same model, same brand language, different formats and aspect ratios.
Fix labels, swap colorways, or extend backgrounds without re-rendering the rest of the image — original details stay intact.
Render dense and small text with high accuracy — useful for translations, multi-region campaigns, and regulated copy.
Generate variant after variant with predictable changes — perfect for A/B tests, mood exploration, and design reviews.
- 1
Select GPT Image 1.5
Open the playground above and pick gpt-image-1.5 while it is still available (until December 1, 2026). Free credits land in your account at sign up — no payment required to start.
- 2
Describe the result you want
Use plain English or Chinese. Mention style, lighting, composition, and any on-image text. Reference real photography terms if you know them.
- 3
Generate or edit
For new images, just generate. For edits, upload a reference and mask the region to change — 1.5 will preserve everything outside the mask.
- 4
Iterate fast
Tweak the prompt, change the size, or seed a new variation. 1.5 is 4× faster than 1.0, so iteration is cheap and immediate.
| GPT Image 1 | GPT Image 1.5 | |
|---|---|---|
| Speed | Standard | 4× faster |
| Edit fidelity | May re-render entire image | Preserves details precisely |
| Text rendering | Good | Best in class |
| Consistency | Variable across generations | Consistent style and composition |
| Complex scenes | Can struggle | Handles reliably |
| Price (output) | $40/1M tokens | $32/1M tokens |
Sign in above to claim your free credits. GPT Image 1.5 still works here until December 1, 2026; the same credits run GPT Image 2 and GPT Image 2.5, which is where new projects should start.