Fei-Fei Li Global Laboratories is accelerating the global prototype race with its first commercial product, Marble

Laboratories of the worlda startup founded by an artificial intelligence pioneer fi fi li, Launches its first global commercial prototype product. marble Available now across free and paid tiers that allow users to convert text prompts, images, videos, 3D layouts, or panoramas into editable and downloadable 3D environments.

Launch of the global generative model First released in limited beta Previewed two months ago, it comes just over a year after World Labs He came out of hiding With $230 million in funding, it puts the startup ahead of competitors building global models. Global models are artificial intelligence systems that generate an internal representation of the environment, and can be used to predict future outcomes and plan actions.

Start-up companies such as Descartes and Odyssey They have released free demos, and Google Genie It is still in limited research preview. Marble is different from this — and even World Labs’ real-time model, rtfm – Because it creates static, downloadable 3D environments instead of creating worlds on the fly while exploring. The company says this results in less shifting or inconsistency, and allows users to export worlds as patches, meshes, or Gaussian videos.

Marble is also the first model of its kind to offer native AI-based editing tools and a hybrid 3D editor that allows users to block out spatial structures before the AI ​​fills in the visual details.

“This is a whole new class of models that generate 3D worlds, and this is something that will improve over time. It’s something we’ve already improved a lot,” Justin Johnson, co-founder of World Labs, told TechCrunch.

last december, World Labs showed how its early models Interactive 3D scenes can be created based on a single image. Although impressive, the somewhat cartoonish scenes weren’t fully explorable since the movements were limited to a small area, and there were occasional glitches in the presentation.

In my beta preview, I found that Marble created impressive worlds with image prompts alone—from game-like environments to real-life versions of my living room. The sights shifted at the edges, although this was apparently improved in today’s launch. However, the world I created in beta with a single prompt looked better and aligned with my goal more closely than the same prompt now.

I haven’t tested the editing features yet, though Johnson says they make Marble practical for gaming, visual effects, and virtual reality (VR) projects in the near term.

“One of our main themes moving forward at Marble is creative control,” Johnson said. “There should always be a fast track to creating something, but you should be able to dive deeper and have a lot of control over the things you create. You don’t want a machine to take the wheel and take all that creativity away from you.”

Marble’s approach to creative control starts with flexibility of input. The beta version only accepted single images, forcing the model to create invisible details for the 360-degree view. With the full launch, users can now upload multiple images or short clips to show a space from different angles and have the model generate somewhat realistic digital twins.

Then we have Chisel, an experimental 3D editor that lets users block out rough spatial layouts (such as walls, boxes, or planes) and then add text prompts to guide the visual style. Marble creates the world, separating structure from style – similar to how HTML provides the structure of a website and CSS adds colors. Unlike text-based editing, Chisel lets you work directly with objects.

“I can just go there and pick up the 3D block that is the couch and move it somewhere else,” Johnson said.

Another new feature that gives you more editing control is the ability to expand the world.

“Once you create a world, you can expand it once,” Johnson said. “When you move to a part of the world that is starting to break apart, you can basically tell the model to expand there or create more of the world close to where you currently are, and then it can add more detail in that area.”

Users who want to create very large spaces can combine multiple worlds with Composer Mode. Johnson showed me this with two worlds he had already built: a room made of cheese with grape chairs, and another that was a futuristic meeting room in space.

The path to spatial intelligence

Marble is available across four subscription tiers: Free (four generations of text, image, or panorama), Standard ($20 a month, 12 generations plus multiple photo/video input and advanced editing), Pro ($35 a month, 25 generations with scene expansion and commercial rights), and Max ($95 a month, all features and 75 generations).

Johnson believes the initial use cases for Marble will be gaming, visual effects for movies, and virtual reality.

Game developers have mixed feelings about the technology. Modern Game Developers Conference Survey It found that a third of respondents believe that productive AI is having a negative impact on the gaming industry – 12% more than the survey indicated the previous year. Intellectual property theft, power consumption and reduced quality from AI-generated content were among the top concerns aired. Last year, A Wired The investigation found that game studios like Activision Blizzard are using AI to cut corners and combat attrition.

In gaming, Johnson sees developers using Marble to create background environments and ambient spaces and then importing those assets into game engines like Unity or Unreal Engine to add interactive elements, logic, and code.

“It’s not designed to replace your entire existing pipeline of games, but just to provide you with assets that you can add to that pipeline,” he said.

For visual effects work, Marble avoids the inconsistency and poor camera control that plagues AI video generators, according to Johnson. Its 3D assets allow artists to render scenes and control camera movements with perfect frame accuracy, he said.

While Johnson said World Labs isn’t focused on virtual reality (VR) applications at the moment, he noted that the industry is “hungry for content” and excited About launch. Marble is already compatible with Vision Pro and Quest 3 VR headsets, and each world created can be viewed in VR today.

Marble may also have potential use cases for robotics. Johnson noted that unlike image and video generation, robots do not have the advantage of a large repository of training data. But with generators like Marble, it becomes easier to simulate training environments.

According to A Final statement By Fei-Fei Li, CEO and Co-Founder of World Labs, Marble represents the first step towards creating a “truly spatially intelligent global model.”

Lee believes that “the next generation of global models will enable machines to achieve spatial intelligence at a completely new level.” If large language models can teach machines to read and write, Lee hopes that systems like Marble can teach them to see and construct. She says the ability to understand how objects exist and interact in 3D spaces could eventually help machines achieve breakthroughs beyond gaming and robotics, and even into science and medicine.

“Our dreams of truly intelligent machines would not be complete without spatial intelligence,” Lee wrote.

Got sensitive advice or confidential documents? We report on the inner workings of the AI ​​industry – from the companies shaping its future to the people affected by its decisions. Connect with Rebecca Bellan on rebecca.bellan@techcrunch.com Or email Brandom at russell.brandom@techcrunch.com. For secure communication, you can contact them via Signal at @rebeccabellan.491 And Russell Brandom.49.