Text to image
Elo · text-to-image arena
GPT Image 2 overview
GPT Image 2 by OpenAI
The model combines advanced multimodal training with diffusion-based image generation. This enables it to convert complex instructions into visually consistent outputs while preserving strong control over composition, typography, and layout.
GPT Image 2 specs
- Model ID
openai/gpt-image-2- Developer
- OpenAI
- Input
- Text, Image
- Output
- Image
- Endpoint
POST https://api.aimlapi.com/v1/images/generations
What Is GPT Image 2 API?
GPT Image 2 is OpenAI's third-generation flagship image model, officially launched on April 21, 2026. It follows gpt-image-1 (March 2025) and gpt-image-1.5 (December 2025), representing the most significant architectural leap in the series.
What sets GPT Image 2 apart from everything before it is a fundamental shift in how the model approaches generation. Rather than jumping straight from a text prompt to pixels, GPT Image 2 first thinks. It reasons about composition, structure, and accuracy before committing to an output. That reasoning step, borrowed from OpenAI's O-series language models, is what makes it the industry's first truly agentic image generation model.
Core Capabilities
GPT Image 2 does not improve any single dimension of image generation, it expands what the category can do. These are the capabilities that matter most for real production workflows.
Agentic Reasoning
Before generating a single pixel, the model researches, plans, and reasons about image structure. This is the first image model with O-series reasoning built in, resulting in fewer failed generations on complex briefs.
Built-In Web Search
GPT Image 2 can query the web in real time before generating, confirming brand logos, event details, product designs, and geographic references that would otherwise be approximated or hallucinated.
Near-Perfect Text Rendering
Typography inside generated images now reads correctly more than 99% of the time. Multi-line headlines, CTA buttons, UI labels, and fine-print captions are all handled reliably, including mixed-script layouts.
2K Resolution & Flexible Aspect Ratios
Outputs up to 2048px, with aspect ratios from 3:1 (ultra-wide banners) to 1:3 (mobile screens). Covers every production format from social ads to presentation slides without post-processing resizing.
Use cases
GPT Image 2 is documented as suited to producing marketing and brand assets such as product mockups, ad creatives, and social graphics with accurate text and logo placement. It is also suited to UI and product design work, including realistic interface screenshots and app mockups for early-stage prototyping and stakeholder presentations. Another documented use case is publishing and editorial work, including book covers, infographics, and illustrated articles where readable text inside images is required. It is also described as useful for e-commerce tasks such as product photography variations, background replacements, and packaging visualization.
GPT Image 2 benchmarks by 
Artificial Analysis
Human-preference results for GPT Image 2 in text to image, next to models available on AI/ML API. Data retrieved Nov 3, 2026.
Models on AI/ML API
Arena results for models you can call with the same API key.
Text to image
Blind pairwise human votes on images generated from text prompts. GPT Image 2 ranks #3 of 98.
GPT Image 2 API pricing
| Usage | Price | Unit |
|---|---|---|
| Input | $5 | 1M tokens |
| Output | $30 | 1M tokens |
| Cached input | $1.25 | 1M tokens |
Prices from the AI/ML API price list, checked on Nov 4, 2026.
GPT Image 2 vs other models: token prices
AI/ML API input and output rates for 5 models, per million tokens.
| Model | Input | Output |
|---|---|---|
 GPT Image 2 | $5 | $30 |
| GPT Image 1 Mini | $2 | $8 |
| GPT Image 2.5 Flare | $5 | $30 |
| GPT Image 2.5 Sunburst | $5 | $30 |
| GPT Image 1.5 | $5 | $32 |
How to use the GPT Image 2 API
View docs
Build with your agent
Give your agent the model details, code and documentation.
Paste it into your coding agent to build the integration.
1Get your API key
Create API keyexport AIMLAPI_KEY="8f2b4c7d9a1e6b3f…"2Make your first request
curl 'https://api.aimlapi.com/v1/images/generations' \
-H "Authorization: Bearer $AIMLAPI_KEY" \
-H 'Content-Type: application/json' \
-d '{
"model": "openai/gpt-image-2",
"prompt": "A ginger cat asleep on a stack of exactly three hardcover books in a rainy bookshop window at dusk. Gold lettering on the glass reads OPEN LATE. Warm lamp light inside, raindrops and blurred neon reflections outside, 35mm film photo."
}'import os
import requests
headers = {
"Authorization": f"Bearer {os.environ['AIMLAPI_KEY']}",
"Content-Type": "application/json",
}
payload = {
"model": "openai/gpt-image-2",
"prompt": "A ginger cat asleep on a stack of exactly three hardcover books in a rainy bookshop window at dusk. Gold lettering on the glass reads OPEN LATE. Warm lamp light inside, raindrops and blurred neon reflections outside, 35mm film photo."
}
response = requests.post("https://api.aimlapi.com/v1/images/generations", headers=headers, json=payload, timeout=120)
response.raise_for_status()
print(response.json())const response = await fetch("https://api.aimlapi.com/v1/images/generations", {
method: "POST",
headers: {
Authorization: `Bearer ${process.env.AIMLAPI_KEY}`,
"Content-Type": "application/json",
},
body: JSON.stringify({
"model": "openai/gpt-image-2",
"prompt": "A ginger cat asleep on a stack of exactly three hardcover books in a rainy bookshop window at dusk. Gold lettering on the glass reads OPEN LATE. Warm lamp light inside, raindrops and blurred neon reflections outside, 35mm film photo."
}),
});
if (!response.ok) throw new Error(`HTTP ${response.status}`);
console.log(await response.json());Models similar to GPT Image 2
.svg){kind=small}
Related articles
.webp)
GPT Image 2 FAQ
What inputs and outputs does GPT Image 2 support?
GPT Image 2 takes image, text as input and returns image.
How much does GPT Image 2 cost?
GPT Image 2 costs $5 input and $30 output per 1M tokens on AI/ML API, billed as you go with no subscription.
How is GPT Image 2 billed?
GPT Image 2 is billed per token, charged separately for input and output.
Can GPT Image 2 take an existing image as input?
Yes, GPT Image 2 accepts image input as well as a text prompt.
Who created GPT Image 2?
GPT Image 2 was built by OpenAI.
How do I call GPT Image 2 on AI/ML API?
Send a request to https://api.aimlapi.com/v1/images/generations with openai/gpt-image-2 as the model id.
Is GPT Image 2 available through a single API key?
Yes. GPT Image 2 is served through AI/ML API, so the same key and endpoint format used for other models applies.
What is the GPT Image 2 model ID?
Use openai/gpt-image-2 as the model parameter. Requests go to https://api.aimlapi.com/v1/images/generations with your AI/ML API key.