OpenAI intros faster, multimodal model GPT-4o; unveils ChatGPT app for Mac

OpenAI has introduced its latest GPT-4o, a significant upgrade to its GPT-4 language model powering ChatGPT. This new iteration — “O” that means “omni” — boasts significant improvements in speed and, crucially, the ability to handle text, vision, and audio.

This comes with a free tier with basic functionalities, while a paid option provides users with five times the capacity. The rollout will be phased, with text and image capabilities arriving in ChatGPT first.

The key highlight of GPT-4o lies in its “native multimodality.” Users can interact with it through text, images, video, or even voice commands, paving the way for advanced voice assistant experiences in future versions of ChatGPT.

“It can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time in a conversation,” OpenAI stated in its official webpage.

This short demo TikTok Shop → from OpenAI’s TikTok account showcases ChatGPT app being able to successfully understand the employee’s surroundings through the iPhone’s camera.

Furthermore, OpenAI will also be releasing a desktop app ChatGPT for Mac with plans to release a Windows version later this year. Rollout has started on Monday with ChatGPT Plus subscribers first and goes for a wider launch in the coming weeks.

In OpenAI’s full demo of GPT-4o, ChatGPT was able to understand a whole chain of code and even understand a graph when given access to the computer screen.

All these while using Voice Mode that basically allows users to talk with ChatGPT directly on the Mac.

This news came shortly after a previous report claiming that OpenAI has entered an agreement with Apple to integrate ChatGPT on the iPhone with iOS 18 taking precedence.