DALL-E is retired — gpt-image-2 is OpenAI's image generator now, built right into ChatGPT.
DALL-E was OpenAI's original image generation line, and it is now retired: the dall-e-2 and dall-e-3 API models were shut down on 2026-05-12. Its successor is gpt-image-2, OpenAI's current image model, alongside gpt-image-1.5, gpt-image-1, and gpt-image-1-mini — available within ChatGPT and via the OpenAI Image API with per-token pricing rather than the old per-image pricing. The strengths that made DALL-E popular carry forward: following complex prompts accurately, rendering readable text within images, and iterating conversationally inside ChatGPT. PxlPeak integrates OpenAI image generation into content production workflows via the API for automated image generation at scale, and migrates legacy DALL-E pipelines to gpt-image-2.
2026-05-12
Date dall-e-2 and dall-e-3 were shut down (OpenAI)
$8/1M
gpt-image-2 input token price (OpenAI)
$30/1M
gpt-image-2 output token price (OpenAI)
gpt-image-2 is OpenAI's current image model (DALL-E 2 and 3 retired 2026-05-12)
Native ChatGPT integration for conversational image creation
Strong text rendering within generated images
Accurate adherence to complex, multi-element prompts
API access with per-token pricing for automated pipelines
Smaller gpt-image-1.5, gpt-image-1, and gpt-image-1-mini models for cost-sensitive workloads
Generate blog post featured images directly from article briefs
Create social media graphics with text overlays in ChatGPT
Build automated visual content pipelines via the Image API
Migrate legacy DALL-E API integrations to gpt-image-2
Assess
We analyze your business needs and how OpenAI Image Generation (formerly DALL-E) fits into your workflow.
Configure
Set up OpenAI Image Generation (formerly DALL-E) with custom settings, integrations, and data connections.
Integrate
Connect to your existing tools — CRM, helpdesk, email, and more.
Train & Launch
Train your team, document everything, and provide ongoing support.
You need the highest artistic quality and don't mind a learning curve — Midjourney produces more aesthetically striking images out of the box.
You need unlimited free generation at scale — self-hosted Stable Diffusion or Flux eliminates per-image costs entirely.
You require fine-tuning on proprietary data — Stable Diffusion is open-source and supports LoRA/DreamBooth fine-tuning; OpenAI's image models do not.
You need precise control over composition and style — Midjourney's --sref and ControlNet in Stable Diffusion offer more control than OpenAI's image models.
Blog Image Automation
OpenAI Image API (gpt-image-2) + ChatGPT + WordPress + n8n
New blog post triggers n8n workflow → ChatGPT generates image prompt from article summary → gpt-image-2 creates featured image → auto-published to WordPress with alt text.
Product Mockup Generator
OpenAI Image API (gpt-image-2) + Shopify + Airtable + Cloudinary
Product descriptions in Airtable trigger gpt-image-2 mockup generation → images optimized via Cloudinary → uploaded to Shopify product listings automatically.
Ad Creative Pipeline
OpenAI Image API (gpt-image-2) + ChatGPT + Canva + Meta Ads API
ChatGPT generates ad copy variations → gpt-image-2 creates matching visuals → assembled in Canva templates → deployed to Meta Ads for A/B testing.
Presentation Assistant
gpt-image-2 + ChatGPT + Google Slides API
Paste slide text → ChatGPT generates image prompts → gpt-image-2 creates relevant visuals → auto-inserted into presentation with layout optimization.
Text in images occasionally misspelled or garbled
OpenAI's image models are strong at text but not perfect. Always review text rendering before publishing. We add a QA step to automated pipelines for text-heavy images.
API rate limits during bulk generation campaigns
OpenAI rate limits apply to Image API calls. We implement queuing, retry logic, and spread generation across off-peak hours for bulk campaigns.
Generated images too generic for brand differentiation
Without style constraints, the model produces a default 'clean digital art' look. We develop detailed prompt templates with brand-specific style language to differentiate.
Content policy rejections on legitimate business prompts
OpenAI's content filters can be aggressive with people, medical, and some product images. We develop prompt workarounds and have fallback workflows using Midjourney or Stable Diffusion.
Ensure your team has a paid ChatGPT plan for interactive use or set up the OpenAI Image API (gpt-image-2) for programmatic access
If you have a legacy DALL-E integration, migrate it — dall-e-2 and dall-e-3 were shut down 2026-05-12
Create prompt templates for your most common image types: blog headers, social posts, product concepts, presentations
Set up an n8n/Zapier workflow for any automated image generation pipeline
Establish a QA process for text-heavy images (the models are good but not perfect at text rendering)
Configure API rate limits and spending caps if using programmatic generation — billing is per token, not per image
Build a library of successful prompts organized by use case for team reuse
Train the team on natural language prompting — the models respond to conversational descriptions better than keyword lists
Set up fallback workflows: if OpenAI's content filter rejects a prompt, route to Midjourney or Stable Diffusion
OpenAI image generation (formerly DALL-E — retired 2026-05-12, replaced by gpt-image-2) is the most accessible option for businesses: it's built into ChatGPT, has a clean API, and OpenAI's content policies mean generated images are commercially safe. It's not the most artistic (Midjourney wins there) or the most photorealistic (Flux wins there), but it's the easiest to integrate and the safest for business use.
OpenAI API account with Image API (gpt-image-2) access
Brand guidelines: colors, styles, visual standards
Use case defined: product shots, marketing assets, social content, or UI mockups
Set up API access
1 dayConfigure OpenAI API keys with appropriate rate limits and spending caps. gpt-image-2 is billed per token: $8 per 1M input tokens plus $30 per 1M output tokens.
Develop prompt templates
2-3 daysCreate reusable prompt templates for your common image types: social media posts, blog headers, product mockups. Include brand style guidelines in every prompt.
gpt-image-2 works best with detailed, descriptive prompts. Vague prompts produce generic images. Be specific about style, composition, and mood.
Build generation pipeline
2-3 daysIntegrate the OpenAI Image API into your content workflow. Automate batch generation for campaigns. Add approval step for brand consistency.
Deploy with monitoring
1 dayTrack image generation costs, team usage, and quality satisfaction. Set up a library of approved prompts.
Expecting Midjourney-level aesthetics
OpenAI's image models prioritize accuracy and safety over artistic style. For premium creative work, pair them with Midjourney or use gpt-image-2 for iterations and ideation.
Not including style direction in prompts
Without style guidance, the model defaults to generic looks. Always specify: photorealistic, illustration, minimalist, etc.
Still calling the DALL-E API
dall-e-2 and dall-e-3 were shut down on 2026-05-12 — any pipeline still calling them fails. Migrate to gpt-image-2 and re-check cost assumptions: pricing is per token now, not per image.
Image generation in ChatGPT can iterate conversationally. Describe what to change and it adjusts — much faster than re-prompting from scratch.
Use the edit API for inpainting: modify specific parts of an existing image while keeping the rest intact. Great for product placement mockups.
┌──────────────┐ ┌──────────────┐ ┌──────────────┐
│ Prompt │────▶│ OpenAI │────▶│ Generated │
│ Template │ │ gpt-image-2 │ │ Image │
│ + Brand │ │ API │ │ (PNG/URL) │
└──────────────┘ └──────────────┘ └──────┬───────┘
│
┌──────────────────────┼──────────┐
▼ ▼ ▼
┌──────────────┐ ┌──────────┐ ┌──────────┐
│ Asset │ │ CMS │ │ Social │
│ Library │ │ Upload │ │ Publish │
└──────────────┘ └──────────┘ └──────────┘// gpt-image-2 — Image generation with brand guidelines
// (dall-e-3 was shut down 2026-05-12; billing is per token now)
const response = await openai.images.generate({
model: "gpt-image-2",
prompt: "A clean, minimalist product photograph of [product] on a white background, soft studio lighting, professional commercial photography style, brand colors: [hex codes]",
n: 1,
size: "1024x1024"
});Social Media Content Pipeline
ChatGPT generates social post copy + image prompt. gpt-image-2 creates the visual. n8n orchestrates the workflow and pushes to Buffer for scheduled publishing across platforms.
Blog Header Automation
When a blog post is published, n8n generates a custom header image via gpt-image-2 using the post title and category. Image auto-uploaded to WordPress media library and set as featured image.
Want us to handle the implementation?
Our team handles OpenAI Image Generation (formerly DALL-E) setup, integration, training, and ongoing support.
Get OpenAI Image Generation (formerly DALL-E) ImplementedMidjourney vs DALL-E
Midjourney wins by default — DALL-E was retired May 12, 2026 (successor: gpt-image-2)
DALL-E vs Stable Diffusion
Stable Diffusion wins by default — DALL-E was retired May 12, 2026 (successor: gpt-image-2)
DALL-E vs Flux
Flux wins by default — DALL-E was retired May 12, 2026 (successor: gpt-image-2)
OpenAI shut down the dall-e-2 and dall-e-3 API models on 2026-05-12. Image generation in ChatGPT and the API now runs on the gpt-image family — gpt-image-2 is the current model, with gpt-image-1.5, gpt-image-1, and gpt-image-1-mini also available. If you have a pipeline still calling the DALL-E API, it no longer works; PxlPeak migrates those integrations to gpt-image-2.
OpenAI's gpt-image models are strong at following complex prompts accurately and rendering text in images. Midjourney produces higher artistic quality and more visually striking outputs. PxlPeak recommends OpenAI image generation for teams that need accuracy and text, Midjourney for teams that need aesthetic quality.
Yes. The OpenAI Image API supports programmatic image generation at scale with gpt-image-2, billed per token ($8 per 1M input tokens, $30 per 1M output tokens) rather than per image. PxlPeak builds automated pipelines that generate images from structured data — for example, creating product images from catalog descriptions or blog illustrations from article summaries.
Yes. OpenAI grants full commercial rights for images generated through ChatGPT and the API. There are no additional licensing fees.
If your team already uses ChatGPT, image generation is available immediately. PxlPeak adds value by building API integrations, prompt templates, and automated workflows — typically delivered in about 1 week.
Midjourney
The gold standard in AI art — photorealism, illustration, and everything between.
Stable Diffusion
Open-source image AI — maximum control, zero per-image cost.
Flux
The photorealism king — 12 billion parameters of visual fidelity.
Adobe Firefly
Enterprise-safe AI images — commercially licensed from day one.
Replace manual workflows with agentic AI ecosystems that pay for themselves.
Ready?
Call now and talk to Aria, our AI strategist — or book a free 30-minute assessment.
Aria picks up instantly · 24/7 · Free assessment · 30-day guarantee
