DALL-E 3 Is Being Deprecated May 12 — Here's Your Migration Guide

On May 12, 2026, OpenAI is pulling the plug on DALL-E 2 and DALL-E 3. DALL-E 2 is already gone from the API as of January 2025. DALL-E 3 follows in May. The replacement — GPT Image 1 — uses a different API format, different pricing model, and different response structure.

If your production pipeline sends requests to model="dall-e-3", it will stop working. Here's exactly what's changing, what breaks, and how to migrate in under 5 minutes.

What's Being Deprecated

What Breaks

The migration isn't just a model name swap. OpenAI changed the API surface:

• Token-based pricing — DALL-E 3 charged per image ($0.04-0.12 depending on size/quality). GPT Image 1 charges per token, like a language model. A 1024×1024 image at "medium" quality costs ~4,000 output tokens.
• New quality tiers — DALL-E 3 had "standard" and "hd". GPT Image 1 uses "low", "medium", and "high" — and they map differently.
• Response format changes — The response includes a revised_prompt field (ChatGPT-style prompt rewriting) and different base64 encoding behavior.
• No more style parameter — DALL-E 3's "vivid" and "natural" styles are gone. GPT Image 1 handles style through the prompt itself.

If you have automation, SaaS integrations, or n8n/Make workflows that call dall-e-3 — they will return 404 errors after May 12.

Option 1: Migrate to GPT Image 1 Directly

If you want to stay on OpenAI, here's the minimum change:

# Before (DALL-E 3) — STOPS WORKING May 12
response = client.images.generate(
    model="dall-e-3",
    prompt="A serene mountain landscape at sunset",
    size="1024x1024",
    quality="hd",
    style="natural"
)

# After (GPT Image 1) — OpenAI direct
response = client.images.generate(
    model="gpt-image-1",
    prompt="A serene mountain landscape at sunset",
    size="1024x1024",
    quality="high"
    # No more 'style' parameter
)

Gotchas with this approach:

• Pricing changes significantly — budget for 2-5x cost increase depending on resolution
• You're still locked to a single provider (OpenAI)
• No fallback if OpenAI has an outage
• Rate limits are more restrictive on image endpoints

Option 2: Drop-In Migration via CreativeAI (2 Lines)

CreativeAI's API is fully OpenAI-compatible. Your existing DALL-E 3 code works — including the model name. We accept model="dall-e-3" and automatically route it to the best available image model (GPT Image 1) with no code changes beyond the base URL.

from openai import OpenAI

# Before: OpenAI direct
client = OpenAI(api_key="sk-your-openai-key")

# After: CreativeAI (change 2 lines, keep EVERYTHING else)
client = OpenAI(
    base_url="https://api.creativeai.run/v1",
    api_key="your-creativeai-key"
)

# This STILL WORKS — even after May 12
response = client.images.generate(
    model="dall-e-3",          # We accept this!
    prompt="A serene mountain landscape at sunset",
    size="1024x1024",
    quality="hd"
)

image_url = response.data[0].url

Why this works: CreativeAI translates legacy DALL-E 3 requests into the optimal format for GPT Image 1 (or whichever model delivers the best result). Your code doesn't change. Your response format stays the same. Your pipeline keeps running.

Node.js / TypeScript

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.creativeai.run/v1",
  apiKey: "your-creativeai-key",
});

const response = await client.images.generate({
  model: "dall-e-3",  // or "gpt-image-1" for explicit routing
  prompt: "Product photo of wireless earbuds on marble",
  size: "1024x1024",
});

console.log(response.data[0].url);

cURL

curl -X POST https://api.creativeai.run/v1/images/generations \
  -H "Authorization: Bearer $CREATIVEAI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "dall-e-3",
    "prompt": "Product photo of wireless earbuds on marble",
    "size": "1024x1024",
    "quality": "hd"
  }'

Why Not Just Use GPT Image 1 Directly?

You absolutely can — and CreativeAI supports model="gpt-image-1" natively too. But here's what you get with CreativeAI on top:

For n8n / Make / Zapier Users

If you're using DALL-E 3 through an automation platform, migration is even simpler:

1. Find your OpenAI HTTP node or integration
2. Change the base URL from https://api.openai.com to https://api.creativeai.run
3. Update the API key to your CreativeAI key
4. Done — your workflow keeps running, even after May 12

We also offer a dedicated n8n community node if you prefer a native integration.

Migration Timeline

Don't wait until May 11. The developers who migrated early from Gemini 3 Pro had zero downtime. The ones who waited scrambled for 48 hours. Use promo code DALLE1000 for 3,000 free credits to test your migration today.