Reve

Video in Reve is here. Create beautiful images in Reve and seamlessly bring them to life with video, powered by Veo 3.1. You’ll unlock a new dimension of creative expression and stay in flow. 1/ Text-to-Video Start with just a prompt. Describe what you want to see, and Reve generates the video. 2/ Image-to-Video Upload a start frame, and optionally add an end frame. Veo 3.1 automatically fills in everything in between, animating your images with smooth, natural motion. 3/ Image Reference Combine prompts with image references for highly specific results. You can create your video by generating a new reference image, uploading an image, or choosing from previously attached images, then simply describe what you want to see. Reve brings the elements together seamlessly. Ready to get started? Head to reve.com, click the filmstrip icon, and bring your first image to life. We can't wait to see what you create.

Calling all #ICCV2025 attendees! We're hosting a Reve social where you can meet our team and unwind on Tuesday evening. RSVP here, limited spots remaining: https://luma.com/9dy5x3j4

Let’s get spooky. See what you should be for Halloween this year 🎃 reve.com/halloween

In less than half a year, we went from a small research preview to one of the best image editing models in the world — and, I believe, the best image editing product. Our team is 1/1000th the size of Google/ByteDance, but that allows us to be hyper-focused on one thing: building the best creativity tools in the world.

We landed at 3rd place on LMArena. We are grateful of what we have achieved, training a model from ground up with such a small team of 10 ML researchers and engineers in a span of 5 months. What's more, our primary focus has not been text-based image editing! We are building a new unified model that supports both verbal and non-verbal controls. Our next model comes with a new architecture and is trained with 5x compute. Stay tuned for what's coming next!

We’re focused on building the best end-to-end image editing experience at Reve. Proud to be one of the best editing models in the world on Artificial Analysis. With our research team of just 10 people, we're getting better and we're doing it quickly… stay tuned for what’s coming next!

Our first editing model is one of the best in the world. With Reve, it’s easier than ever to actually use it. Our model excels at single and multi-image edits, with the ability to combine multiple reference images into a single output image. In less than half a year, we went from a small research preview to a top 3 image editing model on LMArena, and we’re just getting started!

Come create with Reve in LA! Swing by our creative night on 10/8. Whether you're a creative technologist, filmmaker, or AI enthusiast, connect with our team and other creatives to get product tips and craft stunning visual art. Details here:  https://luma.com/reve25