OpenAI unveiled its new AI model GPT-4o during its Spring Update. According to OpenAI, the model is twice as fast as GPT-4 Turbo and is available to all users. In addition, the company releases ChatGPT on macOS.
According to OpenAI GPT-4o ‘much faster’ and improves the model’s capabilities for text, images and audio. While GPT-4 models were previously only available to paid users, the GPT-4o model will be available for free to all users ‘in the coming weeks’. However, the limits for paying users are five times higher. The ‘o’ in the name stands for Omni, which should indicate the versatility of the model. OpenAI CEO Sam Altman reports on can ‘understand’ speech, text and images. GPT-4o will also be available for the API. The model is half the price of GPT-4 Turbo and supports five times higher limits.
GPT-4o also improves ChatGPT’s Voice Mode, allowing the chatbot to serve even better as a smart assistant. The model can respond in ‘real time’ to user questions. OpenAI writes that the model needs an average of 320 milliseconds to respond. That would be comparable to human reaction time. With GPT-4, ChatGPT still needed an average of 2.8 seconds to respond, according to OpenAI. The live presentation also showed that GPT-4o is able to detect emotion in the user’s voice. It can also generate voices in various emotional styles. The improved voting experience will be available to Plus users next month.
OpenAI will also launch a desktop version of ChatGPT. The desktop app gets a redesigned user interface that navigation and accessibility should be improved. The desktop app is available for both free and paid users, but can only be used on a macOS system for the time being. The Windows version will be released ‘later this year’. ChatGPT has been available as an app for iOS and Android for some time now.
