DALL-E 3

Editor's Review

DALL-E 3 is a legacy product — treat any 2026 "review" as an obituary. OpenAI removed it from the API on May 12, 2026 (it had launched at $0.04/image), replacing it with the GPT Image model family: gpt-image-2 (April 2026) at $2.50/$4.00/$15.00 per million text-input/image-input/image-output tokens, and gpt-image-2.5 variants (September 8, 2026) at $5/$8/$30. Bing Image Creator dropped it too, moving to Microsoft's MAI models. What survives is image generation inside ChatGPT itself — included in every plan, roughly 50 images/day on Plus ($20/month), ~2/day on free. While it lived, DALL-E 3's signature was prompt fidelity: complex multi-element descriptions rendered accurately, and text-in-image legibility was best-in-class for its era. Its curse was the house aesthetic — everything trended "clean and modern," with no style reference or character-consistency features, and it was slower than Midjourney and Stable Diffusion pipelines. Who should care? If you're choosing an image tool in October 2026, don't choose DALL-E 3 — choose its successor (GPT Image inside ChatGPT) or a dedicated tool. Versus Midjourney ($30/month Standard with unlimited relaxed generation): Midjourney wins on artistic and cinematic quality, style references, and brand imagery; the GPT Image line wins on understanding convoluted prompts and conversational iteration. DALL-E 3's real legacy is proving that chat-based image editing was the interface people wanted.