The new Ray3 video for Luma Ai can think before construction

Thinking models are not uncommon in the world of artificial intelligence. Many companies have, including Openai’s GPT-O3 and Google’s Gemini 2.5. But Ai Image and Video Company Luma Ai I just dropped the first video of the male areas, called Ray3, and is now available.

Thinking model is a type of artificial intelligence model that uses more computing time to process requests and can return and verify his answers. Thinking models usually give you better responses, whether it is more details or a lower rate of errors.

For Ray3, this thinking power means that you can create Amnesty International Videos with more complex serials. Usually, videos of artificial intelligence range from 5 to 10 seconds. (This is at least the sweet spot – Long clips tend to get some soon.) So filling your claim with the procedure sequence leaves a big room for error. Ray3 is to spend more time working through claims means that it can deal better with those most advanced scenes.

Do not miss any non -biased technology content and laboratory -based reviews. Add cnet As a favorite Google source.

Luma Ai Amit Jain CEO said that thinking models can do more than translating the text into pixels.

He said in an interview with CNET: “He is able to evaluate and say,” Oh, this is not good, or I need to be better in this way. “

As with Chatbots thinking models, you can see the steps taken by the model while working. The new visual explanatory comments tool shows you what the model does – such as a sign of characters for control and other areas that must be kept as they are. You can also use this function to distinguish tires and highlight the changes you want to make in successive claims.

Other promotions help produce better clips, including the ability to generate HDR 16 -bit, a higher accuracy that gives the clips details and more clear. You can also take advantage of the new draft mode, which allows you to test ideas quickly and create clips with lower accuracy. Jain explained that creating clips in 20 seconds in drafts, then escalating them to the accuracy of high accuracy when they are ready, which takes about two to 5 minutes.

Creating the video is a matter Increased use From artificial intelligence. Several large technology companies released artificial intelligence models during the past year, from Midjourney to Google’s Veo 3. All of these models aim to enhance the construction and modern improvements to generate higher quality, include the sound (for VEO 3) and a higher liability of the temptation of professional creators as well as artificial intelligence lovers. They have professionals Express a number of fears About the media created from artificial intelligence, especially about training and publishing artificial intelligence models. Number Collective lawsuits by artists It was presented against artificial intelligence companies. Luma Ai’s privacy policy He says he can use the information it provides to improve its services.

For more, check out How to write an effective AI image router and AI’s best photo generators.