OpenAI today rolled out GPT-4o, its newest flagship model that is faster than GPT-4 yet maintains the same level of intelligence and builds on its voice, vision and text capabilities, according to the organization’s announcement.

The new model takes understanding of images to a higher level. In its announcement, OpenAI gave this example: “You can now take a picture of a menu in a different language and talk to GPT-4o to translate it, learn about the food’s history and significance, and get recommendations.”

Future features will improve its natural language, real-time voice conversations, and will add functionality to enable ChatGPT conversations over real-time video, OpenAI said, enabling ChatGPT to “see” a live sporting event and the user can ask it questions about such things as the rules of the sport, and more. A new Voice Mode will be rolling out in an alpha release within weeks.