DALL·E now available in beta
OpenAI
OpenAI's DALL·E image generator is finally opening up to a million waitlisted users. It's the first real taste of AI art tools going mainstream, credits and all.
Based on reporting by OpenAI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
DALL·E is leaving the velvet-rope demo phase. OpenAI says it's inviting one million people off its waitlist over the coming weeks, turning what had been a small circle of testers and researchers into something closer to a real product with real users.
The pricing model is simple enough to explain in a sentence. Everyone gets free credits that refresh monthly, and if you burn through those, you can buy more in blocks of 115 generations for $15. That's roughly 13 cents per image, though OpenAI isn't calling it that explicitly — the framing is credits, not per-image cost, which is a pretty standard move for a company still feeling out how people will actually use the thing.
What's notable here isn't the tech itself, which has been publicly known since DALL·E 2 launched earlier this year. It's the scale. A million invites is a genuine mainstream push, not a trickle. OpenAI has spent months watching people generate everything from dog astronauts to unsettling corporate logos, and now it's ready to let the credit-card economics do some of the talking about whether this is a hobby, a tool, or something people will pay for repeatedly.
The monthly free-credit refresh is the interesting bit strategically. It keeps casual users in the funnel without charging them, while nudging power users — the ones who'll actually generate dozens of images a week — toward the paid tiers. That's a freemium playbook lifted straight from SaaS, applied to a technology that, a year ago, most people had never heard of.
My take — AI-written commentary, not fact-checked reporting
I like that OpenAI picked a real price instead of hiding behind
Read more about this at: OpenAI