Luma AI launches Ray2 video generator with better physics

Luma AI made waves with the launch of its AI-generated video creation platform Dream Machine last summer.

Of course, even though that was only seven months ago, the field of AI video has advanced rapidly with the release of several new AI video creation models from competing startups in the US and China, including Runway, Kling, Pika 2.0, OpenAI’s Sora, and Google. Veo 2, MiniMax’s Hailuo, and open source alternatives like Hotshot and Genmo’s Mochi 1, to name a few. Even Luma itself recently updated its Dream Machine platform to include new still image creation and brainstorming boards, and debuted an iOS app.

But the updates keep coming: Today, the San Francisco-based startup launched Ray 2, The latest AI video generation model is now available through the Dream Machine website and mobile apps to paying subscribers (to get started).

The model offers “fast, natural motion and cohesive physics,” according to the co-founder and CEO Amit Jain says on his X accountand was trained using 10 times more compute than the original Luma AI video model, Ray1.

“This leads to a higher success rate of production-ready and usable generations, and makes video storytelling accessible to a lot more people,” he added.

Luma’s Dream Machine web platform Offers a free tier with a maximum number of variable 720p generations each month: Paid plans start at $6.99 per month: from “Lite,” which offers 1080p visuals, to Plus ($20.99 per month), to Unlimited ($66.49 USD per month) and enterprise ($1,672.92 per year).

Leap forward into video gen

Currently, Ray2 is limited to text-to-video, allowing users to write descriptions that are converted into 5- or 10-second videos.

The model can create new videos within seconds, although it can now take minutes at a time due to increasing demand from new users.

Examples shared by Luma and early testers in the Creators Program show the model’s versatility, including a man running through an Antarctic blizzard surrounded by explosions, and a ballerina performing on an ice floe in the Arctic.

Impressively, all the movements in the sample videos look realistic and smooth – and oftentimes, the subjects move much faster and more naturally than videos from competing AI generators, which often appear to be generated in slow motion.

The model can also create realistic versions of surrealist ideas such as Giraffe surfingas user X @JeffSynthesized Proven. “Ray 2 is the real deal,” he said. Written on X.

Other AI video creators who have tried out the new model seem to largely agree Jerrod Law publishes on X: “The cinematography, lighting and realism have all improved and it’s fantastic.”

“…It’s so good!” AI video artist Heather Cooper chimed in.

My own tests were a mixed bag, with some of the more complex claims producing unnatural and confusing results. But when I produced clips that more closely resembled what I had in mind in my prompts – e.g Duelists cross swords on a space station orbiting Jupiter – It was impressive, no doubt.

Luma will also add photo-to-video, video-to-video and editing capabilities to Ray2 in the future, further expanding the tool’s creative possibilities, Jain said.

To celebrate the launch of Ray2, Luma Labs is hosting the Ray2 Awards, giving creators the chance to win prizes of up to $7,000. These include:

1. Wide award: The creator whose Ray2 content receives the most views on one platform within the first week of launch will win $5,000. Applications are due by January 22, 2025.
2. Draw for $3000: Creators can enter by sharing Ray2 content on social media and interacting with the Luma AI launch video. The deadline to enter is also January 22.

Winners of both awards will be announced on January 27. Entries can be uploaded via forms provided by Luma Labs, and creators are encouraged to use the hashtags #Ray2 and #DreamMachine when sharing their work.

Additionally, Luma Labs has launched an affiliate program, allowing participants to earn commissions by promoting its tools.