OpenAI announced the launch of an updated version of GPT-4. GPT-4o (with an “o” at the end) has become smarter and faster. It is a multimodal AI model that understands and can generate text, voice and image. The user can interact with the AI in any of the voiced ways. OpenAI conducted a demonstration of talking to GPT-4o using voice. GPT-4o not only responded almost instantly when the speaker ended the call. The AI also responded by converting text to speech with a full sense of real-time communication.

Many interesting and funny videos have appeared on the OpenAI channel demonstrating the capabilities of the new version of AI. For example, GPT-4o helps people learn Spanish by naming objects they see through the camera, helps them figure out a geometry problem, or acts as a guide for a blind person. GPT-4o changed its voice depending on the questions, from a dramatic speech to a cold, robotic tone.

The demonstration also showed that GPT-4o can act as a voice translator between two people speaking different languages. The new chat gpt model was able to not only explain what the code does, but also tell what would happen if certain parts of it were changed.

OpenAI has decided to allow everyone to use this new technology, although paid users will have many more options. The new technology will be rolled out over the next few weeks. A ChatGPT desktop application with voice and visual capabilities will be rolled out soon.