OpenAI has announced the GPT-4o model, its newest LLM set to launch iteratively over the next few weeks. GPT-4o provides GPT-4 power refined to be much easier to use—the model focuses on much greater speeds and user-friendly features. The update will be released for all users of ChatGPT, paid and free.

GPT-4o will enable ChatGPT to become a legitimate Siri competitor, with real-time conversations via voice that are responded to instantly without lag time. ChatGPT also gains a new "Vision" tool alongside the conversation tool that allows it instant visual input that it can respond to in real-time. Use cases demonstrated include helping with math homework and evaluating facial emotion. The tool will also gain "Memory," a mode allowing it greater access to past queries for smoother conversation.

ChatGPT will also become much more accessible, releasing a new desktop client for PC and Mac. The mobile app and web interface are also gaining smooth new UI upgrades to make ChatGPT as easy to use as possible, a stated part of OpenAI's mission statement. With this increased accessibility, GPT-4o was engineered to be fully featured in 50 languages, covering 97% of internet users, according to OpenAI. The GPT-4o upgrade and desktop client will be released for free to all users, without the paywall that some flagship GPT models have had in the past, and will open up the GPT store to free users as well. OpenAI's user base has now swelled beyond 100 million users.

The heavy Apple product placement at the conference and other factors in the past weeks have contributed to a pervasive rumor that OpenAI and Apple are set to announce a partnership at WWDC in June, though nothing in the update video pointed to this. Nvidia did see a namedrop; however, Jensen Huang was thanked for his donation of the DGX H200 computer in April , which powered the demo.

This is breaking news...more to come.