Is DALL-E being discontinued?

DALL-E has not been formally discontinued, but OpenAI has shifted focus to newer image models inside the same API endpoint. DALL-E 3 remains available through ChatGPT and the API as of 2025. Users should check OpenAI's model deprecation page for current status, since older versions like DALL-E 2 are scheduled for retirement on rolling timelines.

What replaced DALL-E inside ChatGPT?

OpenAI has been rolling out newer image generation capabilities under the gpt-image family, which sits inside the same ChatGPT and API surface as DALL-E. Functionally, users see improved text rendering, better instruction following, and more accurate composition. The branding has softened, so many users now refer to the feature simply as ChatGPT image generation rather than DALL-E specifically.

Is SoulGen free to use?

SoulGen offers limited free access, with most features requiring credits or a subscription. Based on the platform's terms, refunds are available within 14 days for unused subscriptions or credit packs, but not once any credits or images have been used. Plan your test usage carefully before committing to a paid tier.

Can I use SoulGen or DALL-E images commercially?

Both platforms permit commercial use of generated images within their respective terms. DALL-E grants users rights to outputs under OpenAI's usage policy. SoulGen allows commercial use provided content complies with applicable law and does not infringe third-party rights. For client work, save the prompt, output, and terms version as evidence of compliant generation.

Which is better for anime-style characters?

SoulGen is generally stronger for anime and stylised character art because its feature set includes dedicated character creation modes and consistent-face tools like DreamTwin. DALL-E can produce anime-influenced output but is tuned more toward general illustration and photorealism, so consistency across multiple images of the same character requires more prompt engineering.

SoulGen vs DALL-E: A Detailed Comparison for UK Creators

What each platform actually does

The two products sit in overlapping but distinct categories. DALL-E is a text-to-image model from OpenAI, currently surfaced through ChatGPT and the OpenAI API. Its job is narrow: you describe an image, it returns a still picture. SoulGen, operated by Synapse AI Limited from Causeway Bay in Hong Kong (registration number 76954156), bundles several generative tools under one account. These include text-to-image, AI video from images or prompts, image editing, outpainting, and Soul Chat for conversations with AI characters.

That difference matters for buying decisions. If your goal is a single illustration for a blog post, DALL-E inside ChatGPT is enough. If you want a character you can chat with, render in multiple poses, and animate with lip-synced speech, the SoulGen feature set covers more of the pipeline without forcing you to stitch together three tools.

Feature comparison: image, video, and characters

DALL-E currently produces stills. Video generation under the OpenAI umbrella sits with Sora, a separate product. SoulGen integrates video directly: you can take a generated image and animate it, or start from a text prompt. The platform also supports human modelling from reference photos and DreamTwin character creation, which lets you keep a consistent face across multiple generations. According to SoulGen's own product pages, talking characters use realistic lip-sync, which is something DALL-E does not attempt because it is out of scope for a still-image model.

On raw image quality for photorealism and editorial illustration, DALL-E 3 remains strong, particularly for clean compositional prompts and text rendering inside images. SoulGen leans toward stylised portraits, anime, and character art. Neither tool is universally better; the right pick depends on whether your output needs to look like a stock photo, a manga panel, or a chat-ready avatar. For a wider angle on stylised competitors, our SoulGen vs Midjourney breakdown covers the artistic-render side in more detail.

Pricing structure and hidden costs

DALL-E pricing flows through OpenAI. Image generation via the API is billed per image at tiered resolutions, and ChatGPT Plus at around 20 USD per month includes image generation as part of the subscription. SoulGen runs on a credits model layered on top of subscription tiers, with image and video actions consuming different credit amounts. The vertical norm across AI companion and image platforms in 2025 sits at roughly 9.99 USD per month for entry tiers and 29.99 USD per month for premium, with token packs starting around 4.99 USD for 100 credits.

A friend recommended a budget-priced AI app to me earlier this year that advertised 9.99 GBP per week. I tested it for a month starting in February. The base model was thin, and the add-ons for voice and video pushed the real weekly spend closer to 25 GBP. That pattern is common: the headline price is rarely the total price, and SoulGen's credit-on-top-of-subscription model can produce the same effect if you generate video heavily. The research point for buyers is simple: model your expected monthly volume in images and seconds of video before subscribing, then compare credit consumption against the included allowance.

Privacy, data retention, and content policy

Privacy handling is where the two diverge most sharply, and this is the area UK users should weigh carefully under UK GDPR. OpenAI retains API inputs for abuse monitoring (typically 30 days unless zero-retention is granted), and ChatGPT conversations may be used for training unless you opt out in settings. SoulGen states that creations not posted to its community are stored for 7 days and then automatically deleted, while published content stays until you remove it. That short retention window is a privacy plus, but it also means you must download your work promptly or risk losing it.

Both platforms enforce content filtering. DALL-E uses OpenAI's moderation stack, which blocks named living people, explicit sexual content, and graphic violence at the prompt layer. SoulGen filters illegal content, real-person impersonation, and non-consensual themes, with both pre-generation prompt scanning and post-generation review. For commercial work, DALL-E grants users ownership of outputs subject to its usage policy. SoulGen's terms permit commercial use provided the content complies with applicable law and does not infringe third-party rights.

Integrations and developer access

This is DALL-E's strongest territory. The OpenAI API plugs into thousands of tools, from Zapier flows to custom apps, and the documentation is mature. If you are a developer building image generation into a product, DALL-E or its successor models inside the OpenAI image endpoint are the path of least resistance. SoulGen is primarily a hosted web product. You sign up via email or Google, work inside the SoulGen interface, and download finished files. There is no public developer API for general use, which limits its place in automated pipelines but simplifies the experience for non-technical users.

For a closer look at open-source alternatives that sit between these two models, the SoulGen vs Stable Diffusion comparison covers the self-hosted route, which appeals to users who want full control over data and weights.

Who should pick which

Choose DALL-E if your workflow is text-to-image for editorial, marketing, or product illustration; if you already pay for ChatGPT Plus; or if you need API access for automation. Choose SoulGen if you want characters you can chat with, animate, and reuse across scenes, or if anime and stylised portraits sit closer to your output goals than photorealism. Many creators end up using both: DALL-E for clean compositional images, SoulGen for character work and video. A full SoulGen review walks through the interface in detail if you want to test the workflow before committing credits. Users specifically looking for companion-style chat experiences in adjacent regions may also consider Candy AI as a comparison point on the conversational side.

SoulGen vs DALL-E: A Detailed Comparison