DALL-E Alternatives

Seven tools to replace OpenAI's <em>DALL-E</em>, picked by the reason you need one, and when ChatGPT Images is enough.

Midjourney is the best DALL-E alternative in 2026 if you liked DALL-E 3 for its artistic look: the Basic plan costs $10 a month for about 800 images. Ideogram is the pick when your images carry text, Recraft when you need vectors, and Gemini when you just want DALL-E's chat workflow for less, with a free tier and a $4.99 plan. Adobe Firefly suits brands that need a commercial-use statement, Leonardo AI and Krea suit people who want many models in one place. And since OpenAI retired DALL-E from its API and from ChatGPT, staying with OpenAI now means ChatGPT Images, which we cover too.

Romain CochardCEO of Hack'celeration
7alternatives compared5criteria each2026prices checked

The ranking is editorial: a sponsored placement is always labeled on its card, and no affiliate link changes a score.

The honest take

Why people are replacing DALL-E

DALL-E was OpenAI's text-to-image model: you typed a description, it returned a picture. DALL-E 2 and DALL-E 3 were sold through OpenAI's API and built into ChatGPT, where most people met them. In 2026 the story changed. OpenAI removed both models from its API, pointing developers to its GPT Image models (gpt-image-2, gpt-image-1 or gpt-image-1-mini), and retired the official DALL·E GPT from ChatGPT. Inside ChatGPT, images now come from ChatGPT Images. So the question is no longer "is DALL-E good?" but "what do I use now, and is OpenAI's replacement the right one for my work?" Here are the five reasons we see people look elsewhere.

DALL-E itself is gone from the API and from ChatGPT

This is the first reason, and it is not a matter of taste. dall-e-2 and dall-e-3 no longer answer API calls: OpenAI lists them as shut down, with gpt-image-2, gpt-image-1 or gpt-image-1-mini as replacements. In ChatGPT, OpenAI's help center answers the question directly: you can no longer generate images with DALL·E, and you should use ChatGPT Images. Any app, script or Zapier step that still calls a DALL-E model has to be rebuilt either way, which is the right moment to compare options rather than swap model names.

The replacement does not look like DALL-E 3

Many DALL-E 3 fans liked its painterly, illustrative look. On OpenAI's own forum, users in 2025 described the newer model as a step backwards for fine-art painting, with flatter local contrast, and one wrote that DALL-E 3 can do what GPT-Image 1.5 cannot. Others defend the new model as great for realism and text. If your style depended on DALL-E 3, Midjourney is the usual destination: it is built around style, mood and lighting rather than literal instructions.

The OpenAI replacements keep moving too

Switching to an OpenAI image model does not end the migration work. gpt-image-1 was shut down in 2026, and gpt-image-1.5, gpt-image-1-mini and even gpt-image-2 are now marked deprecated, with gpt-image-2.6 Sunburst or Flare as the next replacements. A developer who moved from DALL-E 3 to gpt-image-1 has had to move again within the same year, and whoever picked gpt-image-2 faces a second move by early 2027. Ideogram, Recraft and Krea all expose an API too, so this is the moment to test whether another provider's model and pricing fit better.

ChatGPT Images is slow and hard to budget

OpenAI's help center warns that image generation may take a few minutes, depending on the complexity of your request. In ChatGPT, there is no published per-image price or image count for each plan, and the API bills images in tokens rather than per picture, so a heavy month is hard to forecast. Dedicated tools publish clearer units: about 800 Midjourney images for $10, Ideogram credits at one cent each, Krea's 5,000 units shown as about 200 generations.

You need things DALL-E never did

DALL-E produced one kind of output, a raster image. Teams that need editable vectors for a logo (Recraft), long text on a poster (Ideogram), a stated commercial-use promise from the vendor (Adobe Firefly) or one character kept identical across a series (Gemini's Nano Banana, Leonardo's presets) are better served by a specialist. That is why the best answer is often two tools rather than one: a general generator plus the specialist for your most frequent job.
1
Best for artistic image quality

Midjourney

4.4/5

Midjourney is an independent image model you use on its website or in Discord. If you liked DALL-E 3 for its painterly, illustrative look, this is the closest step up: it is still the reference for mood, lighting and style. The Basic plan costs $10 a month for about 200 image jobs, and since one job returns four images, that is roughly 800 images, a little over one cent each. Standard at $30 a month adds unlimited Relax generations, slower jobs that do not eat your Fast hours. Compared with DALL-E inside ChatGPT, you lose the chat: no conversation, no free tier, and prompts work best when they are short and visual. You gain a style range ChatGPT Images does not match, a web editor that works on your own uploaded images, and full ownership of what you create, even after you cancel. If you want a deeper look before you pay, our Midjourney switching guide covers what people move to when they leave it.

Standout features
  • About 800 images a month on the $10 Basic plan (200 jobs of four images)
  • Unlimited Relax image generations from the $30 Standard plan
  • Web editor that works on images you upload
  • Stealth mode on Pro and Mega keeps work off the public Explore page
+Pros
  • ✓The strongest artistic and cinematic look in this list
  • ✓You own your images and videos even after you cancel
  • ✓Yearly billing takes 20% off every plan
−Cons
  • ✗No free plan: every Midjourney plan is a paid subscription
  • ✗Every image is public on Explore by default below Pro
  • ✗Text inside images is less reliable than Ideogram or ChatGPT Images
Verdict

Pick Midjourney if the look of the image is the job. Pick Ideogram instead if your images carry words, and stay in ChatGPT if you mostly want to edit photos in conversation.

See Midjourney plans →
2
Best for text in images

Ideogram

4.2/5

Ideogram is the alternative for people who left DALL-E because the words in their images came out garbled. Posters, ad mockups, book covers, logos and thumbnails are its home ground, and it renders headlines and short slogans more reliably than most models here. It runs on credits, and the unit is simple: one credit is one US cent of list price. An Ideogram 4.5 image costs 3, 6 or 10 credits depending on the quality you pick, so the 2,400 monthly credits of the Plus plan ($20 a month, or $15 a month billed yearly) buy about 400 images at medium quality, or 240 at high quality. The free plan gives 30 credits a week, but only to accounts created with Google or Apple sign-in, and everything it makes is public. The trick worth knowing: on Plus and Pro, running out of credits does not stop you, since Ideogram's own models keep generating in a slower queue with no limit. Where DALL-E inside ChatGPT lets you talk through changes, Ideogram gives you a Studio with layers, masks and a dedicated text tool, which is faster once you know what you want. It also connects to automation tools: there is an Ideogram integration for Make if you generate visuals in bulk.

Standout features
  • Readable headlines and slogans inside the image
  • Credits priced at one cent each, with the cost shown on the run button
  • Studio with layers, masks and a text tool
  • Vector artwork with Ideogram Vector 1 at 13 to 20 credits an image
+Pros
  • ✓Free plan with 30 credits a week to try the text rendering
  • ✓Failed or timed-out generations are not charged
  • ✓Private generation comes with Plus
−Cons
  • ✗An image hidden by the safety check is still charged
  • ✗Unused monthly credits do not roll over
  • ✗Pure artistic style trails Midjourney
Verdict

Pick Ideogram when the text on the image matters as much as the picture. Pick Midjourney for art without words, and Recraft if the end file has to be an editable vector.

Try Ideogram free →
3
Best for vectors and brand assets

Recraft

4.1/5

Recraft is a design tool built around image generation rather than a chat window. Its own models output real vectors (SVG), not just pictures that look flat, and you can lock a brand style so every icon in a set matches. That is the thing DALL-E never did: a DALL-E image was always a raster file you had to trace by hand. Recraft's paid plans also reach partner models like Nano Banana, GPT Image and Flux from the same canvas. The Basic plan costs $12 a month, or $10 a month billed yearly ($120 a year), for 1,000 credits a month; Pro is $20 a month or $16 billed yearly ($192 a year). If you build with code, the API is priced per image: $0.035 for a V4.1 raster image and $0.08 for a V4.1 vector, so a set of 100 vector icons costs about $8. The free plan is fine for trying the editor, but its images are public, owned by Recraft and not licensed for commercial use, so a client logo has to go on a paid plan.

Standout features
  • Native vector output (SVG) from Recraft models
  • Custom brand styles shared across a team on Teams
  • Partner models such as Nano Banana, GPT Image and Flux on paid plans
  • Per-image API from $0.035, $0.08 for a V4.1 vector
+Pros
  • ✓Vectors you can open and edit in a design tool
  • ✓Images made on a paid plan stay yours and private after you cancel
  • ✓Top-up credits never expire
−Cons
  • ✗Free plan images are public, owned by Recraft and not for commercial use
  • ✗Monthly credits do not roll over
  • ✗More options to learn than a chat window
Verdict

Pick Recraft if your output ends up in a brand kit, a website or a print file. Pick Ideogram for one-off posters with text, and Midjourney for painterly art.

See Recraft plans →
4
Best chat-based replacement

Google Gemini

4.0/5

Gemini is Google's assistant, and it is the closest thing to how most people used DALL-E: you describe an image in a chat, then ask for changes in plain words. Image generation and editing run on Google's Nano Banana models, which are strong at keeping a face or a product the same across several edits. The free tier with a Google account includes image generation and editing. Google AI Plus costs $4.99 a month and gives twice the usage of the free tier, plus 200 Google Flow credits for the creative studio; Google AI Pro costs $19.99 a month with 1,000 Flow credits and 5 TB of storage. The catch is the limit itself: Gemini does not publish a fixed number of images. Your allowance refreshes every 5 hours until you hit a weekly cap, and it shrinks with complex prompts and long chats. Compared with ChatGPT Images, the experience is very similar; the reason to switch is price, the Google bundle, or a preference for Nano Banana's edits. Our comparison of AI chatbots with image generation puts the two side by side, and if Gemini's caps are what bothers you, see our Gemini alternatives. Our full Gemini review covers the assistant beyond images.

Standout features
  • Conversational generation and editing, like DALL-E in ChatGPT
  • Image generation and editing on the free tier
  • Google AI Plus at $4.99 a month doubles free usage
  • 200 to 1,000 Google Flow credits on paid plans
+Pros
  • ✓The cheapest paid plan in this list at $4.99 a month
  • ✓Very good at keeping a subject consistent across edits
  • ✓Storage and Google apps bundled into the same plan
−Cons
  • ✗No published image count: limits refresh every 5 hours up to a weekly cap
  • ✗Limits may change without notice
  • ✗Fewer fine controls than Midjourney or Recraft
Verdict

Pick Gemini if you want DALL-E's chat workflow at a lower price. Pick Midjourney or Ideogram once you need a predictable monthly image count.

Open Gemini →Read our Gemini review
5
Best for commercially safe images

Adobe Firefly

3.9/5

Adobe Firefly is Adobe's AI app for images, video and audio, also built into Photoshop through features like Generative Fill. Its pitch is legal comfort: Adobe states that outputs from its own Firefly models are safe for commercial use, and eligible enterprise customers can get IP indemnification for select outputs. For client work, that statement is often what tips the decision. The Standard plan costs US$9.99 a month for 2,000 generative credits, Pro US$19.99 for 4,000, Pro Plus US$49.99 for 10,000. All paid plans include unlimited standard image and vector generations; the credits pay for premium work like video, audio and partner models from Google and OpenAI. That last point matters if you wanted to keep an OpenAI model: inside Firefly it costs credits, and Adobe makes you responsible for checking whether partner-model content suits your project. Credits reset each billing cycle and do not roll over. We cover the reasons people leave it in our Adobe Firefly alternatives page.

Standout features
  • Firefly model outputs stated safe for commercial use
  • Unlimited standard image and vector generations on every paid plan
  • Generative Fill inside Photoshop
  • Partner models from Google and OpenAI in the same app
+Pros
  • ✓The clearest commercial-use statement in this list
  • ✓Works inside the Creative Cloud files you already edit
  • ✓A free plan with daily generations to start
−Cons
  • ✗Partner models, video and audio burn credits that reset monthly
  • ✗Partner-model outputs are your responsibility, not Adobe's
  • ✗Team plans are annual contracts billed monthly, with a cancellation fee
Verdict

Pick Firefly if you work in Photoshop or need a clean commercial-use story. Pick Midjourney for style, and Gemini for a chat-based workflow from $0 or $4.99 a month.

See Firefly plans →
6
Best for many models and game assets

Leonardo AI

3.8/5

Leonardo AI is a creative platform, bought by Canva in 2024, that puts more than 80 models for images and video, style presets and a real-time canvas behind one subscription. It is popular for game assets, characters and concept art, where you want to swap models and presets quickly. Its unit is the fast token. The free plan gives 150 fast tokens a day; Essential costs $12 a month ($10 a month billed yearly) for 8,500 fast tokens a month, Premium $30 ($24 billed yearly) for 25,000, and Ultimate $60 ($48 billed yearly) for 60,000. From Premium up, image generation on select models (Phoenix, Lucid, FLUX Dev among them) is unlimited at a relaxed pace. Two details most guides skip: the free plan runs on automatic model selection only, while paid plans open 80+ models; and unused paid tokens roll into a token bank capped at three months of allowance (25,500 on Essential). Tokens per image vary with the model and settings, so check the cost shown before you run a batch. The thing to read before you leave DALL-E for Leonardo's free plan is the rights clause: free images are public, and for images that are made public, Leonardo keeps worldwide perpetual rights. Paid users can switch on private mode. Our Leonardo AI switching guide covers the other side of that choice.

Standout features
  • Several image and video models in one app
  • 150 free fast tokens every day
  • Unlimited relaxed image generation on select models from Premium ($30)
  • Real-time canvas and presets, plus a Leonardo AI integration for Make for batch jobs
+Pros
  • ✓A generous daily free allowance
  • ✓Good fit for characters and game assets
  • ✓Yearly billing saves up to 20%
−Cons
  • ✗Free images are public, and Leonardo keeps worldwide perpetual rights on public images
  • ✗Free plan picks the model for you: choosing among 80+ models starts on Essential
  • ✗Busier interface than a chat window
Verdict

Pick Leonardo if you want to try many models and presets for one fixed price. Pick Midjourney for a single top-quality model, and Recraft for brand vectors.

Try Leonardo free →
7
Best free sandbox across models

Krea

3.7/5

Krea is a creative suite that rents many image and video models, Nano Banana among them, and adds real-time generation, upscaling and custom model training. It suits someone who lost DALL-E and does not yet know which model to adopt: you can run the same idea through several engines and keep the one that works. Our list of new AI image generators to watch tells you which recent models are worth a run. Its unit is the compute unit, and the pricing page converts it for you: the Basic plan at $9 a month gives 5,000 units, about 200 Nano Banana 2.1 generations, so roughly 4.5 cents an image. Pro at $42 a month gives 20,000 units, about 803 of those generations. The free plan gives 100 units a day, enough for a few images to compare. Two limits: units on Basic, Pro and Max do not roll over, and extra compute packs expire after 90 days. Commercial licensing comes with the paid plans. If you work at scale with open models, our Flux alternatives page compares the hosts in more depth.

Standout features
  • Many image and video models in one place
  • 100 free units every day
  • LoRA training on your own images
  • Image upscaling up to 22K
+Pros
  • ✓Cheap way to compare models before committing to one
  • ✓Unit cost converted into image counts on the pricing page
  • ✓Commercial license on paid plans
−Cons
  • ✗Units do not roll over on Basic, Pro and Max
  • ✗Extra compute packs expire after 90 days
  • ✗The interface takes time to learn
Verdict

Pick Krea if you want to shop between models before choosing one. Pick Midjourney or Ideogram once you know what you need and want a single, simpler tool.

Try Krea free →
At a glance

7 DALL-E alternatives compared

Scores are our editorial score out of 5, the plain average of five criteria: image quality, prompt and text accuracy, editing and control, value for money, and ease of use. Gemini's score is the one from our Gemini review. Prices are monthly unless marked yearly. The last column says what each tool gives you that DALL-E did not. For a wider ranking beyond DALL-E leavers, see our best AI image generators.

Best forEdge over DALL-EFree planTeam sizeVisit
1MidjourneyBest for artistic image qualityA stronger artistic style, about 800 images for $104.4/5$10/mo—Creators and agenciesVisit →
2IdeogramBest for text in imagesReadable text, credits at one cent each4.2/5Free, Plus $20/mo✓Marketers and SMBsVisit →
3RecraftBest for vectors and brand assetsEditable vectors and locked brand styles4.1/5Free, Basic $12/mo✓Designers and brand teamsVisit →
4Google GeminiBest chat-based replacementThe same chat workflow from $0 or $4.99/mo4.0/5Free, Plus $4.99/mo✓Individuals and small teamsVisit →
5Adobe FireflyBest for commercially safe imagesCommercial-use statement and Photoshop3.9/5Free, Standard US$9.99/mo✓Creative Cloud users and brandsVisit →
6Leonardo AIBest for many models and game assetsMany models and presets, 150 free tokens a day3.8/5Free, Essential $12/mo✓Game artists and hobbyistsVisit →
7KreaBest free sandbox across modelsCompare many models, 100 free units a day3.7/5Free, Basic $9/mo✓Explorers and designersVisit →

Prices in US dollars before tax, read on each vendor's pricing page in 2026.

Is your tool missing from this page? Email us: hello@hackceleration.com

Buyer's guide

How to choose a DALL-E alternative

An AI image generator is two things stacked together. The model is the engine that turns a prompt or a reference image into pixels: Midjourney's own model, Ideogram 4.5, Nano Banana, Firefly Image, Flux, Recraft V4.1, or OpenAI's GPT Image. The app is the studio around it, with its plans, credits, licenses, editor and API. Several studios rent the same models (Nano Banana shows up in Gemini, Krea, Recraft and Ideogram), so after DALL-E the real choice is which studio gives you the images, rights and controls you need for the money.

Each app meters usage in a different unit, and that unit decides your monthly bill. A credit (Ideogram, Recraft, Firefly) is a budget that each action draws from at a different rate. A job or GPU time (Midjourney) is processing time: one image job uses about a minute and returns four images. A token or unit (Leonardo, Krea) works like a credit, with a cost that changes by model. Only a few vendors translate their unit into images for you, so do the sum before you subscribe: monthly units divided by the cost of one image at the quality you will really use.

Before comparing, write down three things: how many finished images you need a month, whether those images carry text or must be vectors, and whether they go into client work. Then demand four answers from each tool: what one image really costs at that quality, whether unused units roll over, whether your images are public by default, and what the license says on the plan you will buy, not on the top plan. If you used DALL-E for marketing, our ranking of the best AI image generators for marketing teams goes further on that use.

Leaving over style

If the look was why you liked DALL-E 3, start with Midjourney. Basic at $10 is enough to judge it; move to Standard at $30 only when you need unlimited Relax generations. If you prefer to stay with a chat, Gemini's Nano Banana is the closest conversational experience.

Leaving over text and layout

Posters, ads and thumbnails with words belong in Ideogram. If the deliverable is a logo or an icon set that a designer will edit, Recraft's vectors save the tracing step DALL-E always required.

Leaving over rights and privacy

Check the plan you will actually buy. Midjourney images are public on Explore below Pro, Recraft's free images are owned by Recraft, and Leonardo keeps worldwide perpetual rights on public images. Adobe Firefly gives the clearest commercial-use statement for its own models.

Leaving the API

Developers rebuilding a DALL-E integration can test Recraft's per-image API ($0.035 a raster image, $0.08 a vector) or OpenAI's GPT Image models. Budget for the next migration: all three models OpenAI first proposed as DALL-E replacements (gpt-image-2, gpt-image-1 and gpt-image-1-mini) are now shut down or deprecated.
  • Count your real monthly images and divide the plan's units by the cost of one image at your quality.
  • Check whether unused credits, Fast hours or units roll over (on most plans here they do not; Leonardo's token bank is the exception).
  • Check whether images are public by default and which plan makes them private.
  • Read the commercial-use terms of the exact plan you will pay for.
  • Download your old DALL-E images from your ChatGPT conversations before you close the account.
FAQ · 8 questions

DALL-E alternatives, your questions

  • Is DALL-E still available?
    No, not as DALL-E. OpenAI shut down dall-e-2 and dall-e-3 in its API in 2026 and retired the official DALL·E GPT from ChatGPT. Image generation in ChatGPT now runs on ChatGPT Images, and custom GPTs with Image Generation enabled in their capabilities use the new model rather than DALL-E.
  • What replaced DALL-E 3 in ChatGPT?
    ChatGPT Images replaced it. It is available on every ChatGPT tier, on the web, iOS and Android, and it can also generate and edit images inside Codex. You can upload a photo, save it as a reference, and call it later in any chat by typing @ followed by the name you gave it.
  • What happened to my old DALL-E images?
    They stay in the ChatGPT conversations where you made them, and ChatGPT also gathers the images it creates under Images, on web and mobile. The trap is cleanup: the only way to delete an image from My images is to delete the conversation it came from, so tidying your history wipes them. Before you tidy your history or close the account, open each image and use Save to download it to your device.
  • Can I use AI-generated images commercially?
    Usually yes on a paid plan, but it depends on the tool and the plan. Midjourney requires the Pro or Mega plan for companies with more than $1,000,000 in yearly gross revenue. Stability AI's community license for Stable Diffusion covers commercial use for organizations under $1M in annual revenue. Always read the terms of the exact plan you buy.
  • Which DALL-E alternative runs on my own computer?
    Stable Diffusion and Flux are the usual answers, since their model weights can be downloaded and run locally with tools like ComfyUI. Stability AI's community license is free for individuals and organizations under $1M in annual revenue. You need a capable graphics card and some setup time, and our Stable Diffusion alternatives page covers the hosted options if that is too much.
  • Is there an AI image generator with unlimited images?
    Yes, but only in a slower lane and only for standard images. Adobe Firefly's paid plans, from US$9.99 a month, include unlimited standard image and vector generations. Midjourney Standard ($30 a month) gives unlimited Relax image generations, Leonardo Premium ($30 a month) unlimited relaxed images on select models, and Ideogram Plus and Pro keep generating with Ideogram's own models in a slower queue once the credits run out. Firefly and Ideogram still charge credits for partner models and video.
  • Do unused image credits roll over to next month?
    On most plans, no. Ideogram's monthly credits reset at each renewal, and so do Adobe Firefly's generative credits and Midjourney's Fast hours. Leonardo is the exception among subscriptions: unused fast tokens go into a token bank, capped at 25,500 on Essential and 180,000 on Ultimate. Bought top-ups also differ: Ideogram top-up credits roll over and Recraft's never expire, while Midjourney's purchased Fast time expires after 60 days and Krea's compute packs after 90 days.
  • Can I still use an OpenAI image model through the API?
    Yes, through the GPT Image models, which OpenAI recommends instead of DALL-E. They are billed per million tokens of image input and output: GPT-Image-2.6 Sunburst and Flare cost $8 per million image input tokens and $30 per million image output tokens. Build on those two: gpt-image-1 has already been shut down, and gpt-image-1.5, gpt-image-1-mini and gpt-image-2 are all marked deprecated.
Hack'celeration Lab

Get the next breakdown in your inbox

No spam. Unsubscribe anytime.