Artificial intelligence startup Luma AI Inc. today announced the debut of a powerful new model for image generation that’s available now through its popular Luma AI Dream Machine platform for creators and marketers.

The AI startup, which provides tools for creating and generating images and videos using simple, text-based prompts, also said it’s launching Dream Machine on the web and on Apple Inc. iOS devices, with a limited free service for those who want to test its capabilities, and four new subscription tiers.

In addition, there’s a new image application programming interface for developers who would like to embed its image creation capabilities into third-party apps.

Luma AI launched Dream Machine in June, targeting creatives and other professionals with the service, which leverages customized and extremely powerful image and text-generation models created by the startup. It has grown to more than 25 million users, the company said.

According to the startup, Dream Machine is all about helping creators express what’s on their minds and free their imagination to create stunning visuals that can be used for web content, marketing materials and more.

One of the most interesting capabilities of Dream Machine has always been its ability to support more descriptive prompts. So users can provide it with really detailed instructions, explaining they want to see a video of “a cute Dalmation puppy running after a ball on the beach at sunset, with small children paddling in the sea in the background, and a couple in the foreground walking hand-in-hand.” The result, according to Luma AI, would be an extremely realistic and detailed video depicting exactly what the user wants to see, in less than two minutes.

In addition, Dream Machine supports image prompts, so users can upload various images depicting the kinds of visual themes, structures and colors they need in their images and videos. It also provides tools for sharing and collaborating on created images and videos.

With today’s updates, Dream Machine is getting a refreshed user interface that’s said to enable more “intuitive” image and video creation, with various new personalization features, so users can edit their videos using tools instead of prompt engineering.

As for the new Photon Image Model, this is said to be a high-performance foundational model for image generation that’s aimed at designers, movie makers, educators, small businesses and visual thinkers. It processes user prompts at much faster speeds — up to 800% faster than comparable models — and the end result is promised to be much higher-quality images than before.

It’s being made available to Dream Machine users and also developers through the new Luma AI Photon Image API. That means it can be integrated into other AI applications, with assurances that all inputs and outputs will remain private and never be used for model training purposes.

The company added that Dream Machine is now available as a free and monthly subscription service, with new choices including a hobbyist tier priced at $9.99, an explorer tier that costs $29.99 per month and a professional tier for $99.99. There’s also an enterprise tier for larger organizations, for which pricing was not disclosed.

Luma AI co-founder and Chief Executive Amit Jain said the startup is delivering the next generation of creative intelligence, going far beyond the capabilities of traditional large language models, with better visuals that are richer in context than before.

“We are building Dream Machine on top of this creative intelligence as a new place for people to express their ideas in a fluid image, video and language medium,” he said. “We combine rigorous scientific research with an ambitious vision to help people think more visually, explore their ideas more effectively and share their stories freely.”

Images: Luma AI