I just tested Luma’s new Ray2 AI video generator – and it’s better than Sora

On January 24, 2025By Littum

Luma Labs has given its popular Dream Machine AI creativity platform a major upgrade, bringing the new Ray2 video model into the system. This is a huge upgrade over the previous Ray 1.6, offering better realism and more natural movement.

Ray2 was announced last year as part of a new partnership with Amazon AWS. It has finally been integrated into Dream Machine, available as the default option when creating a video.

The AI startup describes Ray2 as “a new frontier in video generative models.” To achieve the level of visual and motion realism, they increased the computing power 10x compared to previous models, which has “opened up new freedoms for creative expression and visual storytelling.”

I’ve been testing Lumas Ray 2 since launch and the video generations are very impressive, but it’s slow due to demand with some issues around clips refusing to generate or taking too long to be useful. These are all the same childhood ailments that every platform has when they launch a new model. So I would say that Ray2 is definitely on its way to being one of the best AI video generators out there.

Putting Ray2 to the test

Being built into the Dream Machine already gives the Ray2 a leg up over other video models because of how impressive the Dream Machine is to work with. It makes creating content with AI more collaborative and less about throwing a prompt in the wind and hoping for the best.

Accessing Ray2 is as simple as starting a new board in Dream Machine, selecting Video from the prompt bar, and typing your prompt. The AI handles the rest, showing you two videos and providing the usual customizable interface where you can change elements in the prompt.

Because of the problems I mentioned earlier, I was only able to get about half of the prompts I was actually trying to generate, and because of how slow it was, I couldn’t make use of the recall- and the collaborative features that make Dream Machine so good. Despite that, I was still impressed.

In one example, I asked Luma’s Ray2 to make a video of a knife cutting an onion. Knife skills are something no video model – with the exception of the Google Veo 2 – has been able to do consistently well. While not perfect, the movement was spot on and it cut.

Ray2 is also particularly good at animal movements. I asked it to generate several videos of dogs — including one stretching and another catching butterflies — and it did both very well. There were some elements in the butterfly video that weren’t perfect, but with Dream Machine that can be fixed relatively easily by replying to the original video and indicating what needs to be changed.

When I was able to get videos to generate, it happened extremely quickly. It doesn’t look like Luma had to sacrifice generation speed (something the Dream Machine was famous for) in favor of improving quality. We seem to get performance and speed all in a single model.

Luma Labs Ray2 Final Thoughts

Overall, Ray2 appears to be a significant step forward in generative video. Its jump feels very similar to the jump we saw when Luma first launched the Dream Machine. It’s also slightly better in terms of movement than OpenAI’s flagship Sora model.

It’s not perfect. There are still artifact issues; sometimes the movement doesn’t make sense; and it’s currently just text-to-video (though picture-to-video is coming soon). However, these are all issues that plague every other model I’ve tried, including the Sora, Runway, Kling and Pika.

The biggest takeaway is how quickly AI video is evolving. Being able to generate ten seconds of high definition video that is almost indistinguishable from something filmed with a camera would have been unthinkable two years ago. Today it is common and available from several companies.