DALL-E Image Generation

Quick Answer: DALL-E 3 generates images from text prompts via the Azure OpenAI Images API. Supported sizes: 1024x1024, 1792x1024, 1024x1792. DALL-E automatically revises prompts for detail and safety. Content filtering screens both prompts and generated images.

Generating Images with DALL-E 3

from openai import AzureOpenAI

client = AzureOpenAI(
    api_key="<your-key>",
    api_version="2024-06-01",
    azure_endpoint="https://my-openai.openai.azure.com/"
)

response = client.images.generate(
    model="dall-e-3-deployment",
    prompt="A serene mountain landscape at sunset with a reflection in a lake, digital art style",
    n=1,  # DALL-E 3 only supports n=1
    size="1024x1024",  # or "1792x1024" or "1024x1792"
    quality="standard",  # or "hd" for higher detail
    style="vivid"  # or "natural" for more photographic
)

# Get the image URL
image_url = response.data[0].url
revised_prompt = response.data[0].revised_prompt
print(f"Image URL: {image_url}")
print(f"Revised prompt: {revised_prompt}")

DALL-E 3 Parameters

Prompt Revision

DALL-E 3 automatically rewrites your prompt to add:

• More descriptive detail for better image quality
• Safety-oriented language to avoid harmful content
• Artistic direction when the original prompt is vague

The revised prompt is always returned in the response so you can see what was actually used.

Content Filtering for Images

Responsible AI Restrictions

• Cannot generate photorealistic images of identifiable real people
• Cannot reproduce copyrighted characters or trademarked logos
• Violence, self-harm, sexual, and hate content filters apply
• DALL-E refuses prompts asking for harmful, misleading, or deceptive imagery

On the Exam: Know that DALL-E 3 on Azure includes content filtering on BOTH input prompts AND generated images. Questions may test whether you understand the dual-filter approach and the revised prompt feature.