OpenAI GPT-4o: New 'Free' Model With Improved text, Visual And Audio Capabilities

OpenAI's latest update to its artificial intelligence model can mimic human cadences in its verbal responses and can even try to detect people's moods. GPT-4o, short for omni, will power OpenAI's popular ChatGPT chatbot, and will be available to users, including those who use the free version, in the coming weeks,

Hyderabad: OpenAI’s flagship artificial intelligence tool, ChatGPT, has been upgraded to a more powerful and human-like version, GPT-4o- a version of the GPT-4 model, making it available to all users for free on Monday. The update comes a day ahead of Google’s planned announcement of Gemini, its own AI tool that stands in stark contrast to ChatGPT.

According to the company, the latest model matches GPT-4 Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being much faster and 50% cheaper in the API. GPT-4o is especially better at vision and audio understanding compared to existing models.

"We're very, very excited to bring GPT-4o to all of our free users out there," Chief Technology Officer Mira Murati said at the highly anticipated launch event in San Francisco.

The company has trained a single new model end-to-end across text, vision, and audio, for GPT-4o, meaning that all inputs and outputs are processed by the same neural network.

OpenAI CEO Sam Altman expressed his enthusiasm in a blog post, highlighting the newfound naturalness in conversing with computers, a feat previously considered challenging. "It feels like AI from the movies ... Talking to a computer has never felt really natural for me; now it does," the company CEO said.

During a demonstration with Chief Technology Officer Mira Murati and other executives, the AI bot chatted in real time, adding emotion specifically more drama to its voice as requested. It also helped walk through the steps needed to solve a simple math equation without first spitting out the answer and assisted with a more complex software coding problem on a computer screen.

It also took a stab at extrapolating a person's emotional state by looking at a selfie video of their face (deciding he was happy since he was smiling) and translated English and Italian to show how it could help people who speak different languages have a conversation.

On the other hand, Google plans to hold its I/O developer conference on Tuesday and Wednesday, where it is expected to unveil updates to its own Gemini, its AI model.

Read More