Strategically anticipating Tuesday’s Google I/O presentation, OpenAI unveiled GPT-4o (the o stands for omni), the new version of the GPT-4 model that powers the company’s intelligent assistant. Mira Murati, CTO of OpenAI, revealed during a live stream that GPT-4o is significantly faster and introduces improvements in text, visual and audio capabilities. Murati also confirmed that the new model will be available for free to all users, with paying users continuing to enjoy “up to five times the capacity limits” compared to free users.
OpenAI CEO Sam Altman shared more details about GPT-4o via a post on X, noting that the model is “natively multimodal.” This implies that GPT-4o can generate content and understand commands in speech, text and visual formats. Altman also announced that developers will have access to an API that is half the price and twice as fast as GPT-4 Turbo, providing new opportunities for innovation.
GPT-4o features will be released gradually, as explained in an OpenAI blog post, with text and visual capabilities already available on ChatGPT starting Monday. Additionally, ChatGPT’s voice mode will see the integration of new features, allowing the app to act as a real-time voice assistant, capable of responding to and observing its surroundings. Currently, voice mode only responds to one prompt at a time, limiting what it can hear.
In an analysis of OpenAI’s journey, CEO Sam Altman reflected on the company’s changing vision. Originally focused on creating global benefits, OpenAI faced criticism for not open-sourcing its advanced AI models. Altman suggested that the focus has now shifted towards making these models available to developers via paid APIs, allowing third parties to create innovative applications. “We will program AI and then others will use it to create amazing things that will benefit us all,” he said.
Prior to the launch of GPT-4o, there had been several speculations about OpenAI’s upcoming announcements, including an AI search engine to compete with Google and Perplexity, a voice assistant built into GPT-4, or an entirely new and improved model, GPT -5. However, OpenAI has chosen to present GPT-4o, opening a new phase of development and use of AI technologies just on the eve of Google I/O, where numerous announcements are expected from the Gemini team.
#OpenAI #launches #artificial #intelligence #GPT4o #free
