Luma AI Ray3: The First Reasoning AI Video Model

New Era: World’s “First Reasoning Video Model” Unveiled

Luma AI, known for its AI-powered video generation models, has unveiled a groundbreaking innovation — Ray3, presented as the “first video model capable of reasoning.”

Reasoning AI models already power leading chatbots like ChatGPT and Gemini. But Luma AI’s Ray3 takes this further, marking a new stage in AI-driven video creation. The ambitious platform is now accessible via Dream Machine.

Strategic Partnerships

Backed by Nvidia and Andreessen Horowitz, Luma AI has already secured collaborations with Adobe and leading advertising agencies in the US and UK.

What Sets Ray3 Apart

• The model reviews its own generated scenes, leaving revision notes and creating improved versions.
• Users can save progress and continue later.
• A “draft mode” enables quick, low-cost testing; polished results can then be rendered in 4K HDR.

The key differentiator is reasoning ability. According to Luma AI, Luma AI Ray3 produces more consistent characters, naturally evolving scenes, and physics-based motion. Moreover, Ray3 is the first model capable of generating video in 10, 12, and 16-bit HDR ACES2065-1 EXR format, making it suitable not only for social media but also for big-budget commercials and film production.

This is Ray3. The world’s first reasoning video model, and the first to generate studio-grade HDR. Now with an all-new Draft Mode for rapid iteration in creative workflows, and state of the art physics and consistency. Available now for free in Dream Machine. pic.twitter.com/qm29hkDA14

— Luma AI (@LumaLabsAI) September 18, 2025

Expansion and Adobe Integration

Headquartered in Silicon Valley, Luma AI recently opened a studio in Los Angeles and is preparing to launch an office in New York. This positions the company to work closely with Hollywood and major ad agencies.

Most importantly, Ray3 will be directly integrated into Adobe Firefly, significantly boosting its impact for creative professionals.