OpenAI’s Spring Event unveiled GPT 4o, a new iteration of ChatGPT, which promises to revolutionize AI interactions by working seamlessly across voice, text, and visual processing. This latest model boasts significant improvements, including being twice as fast as GPT 4, 50% less expensive, and having 5 times as many rate limits as the previous ChatGPT 4 Turbo.
The announcement was made in a casual presentation by Chief Technology Officer Mira Murati, who introduced the new features and capabilities of GPT 4o. The presentation was followed by live demonstrations led by research leads Mike Chen and Barret Zoph, showcasing the model’s capabilities.
To enhance user experience, OpenAI is also releasing a new desktop app for Mac, with a Windows version soon to follow. This marks an exciting development in AI technology, making it more accessible and efficient for users across various platforms.
The team demonstrated ChatGPT’s new interaction style in real-time, showcasing a more conversational and intuitive experience. They interrupted the chatbot and requested tone-of-voice adjustments, which were seamlessly implemented. The voice output sounded more natural, making it feel like a regular conversation with the video hosts.
GPT 4o’s visual interaction capabilities were also displayed, as the team used a phone camera to point at math problems and sentences. The chatbot successfully interpreted and assisted with the math, even recognizing a heart symbol as “love.” The most impressive aspect was the natural cadence of GPT 4o’s voice, which, although still computer-generated, made it easier to interact with.
The team then moved on to coding examples, where GPT 4o interpreted code, offered improvements, and even ran the code to interpret its output. The demo showed GPT 4o accurately interpreting the code, and the suggested improvements worked well. The hosts could interrupt the chatbot when they no longer needed to hear its explanations, making the interaction feel more dynamic.
Overall, the demonstration showcased GPT 4o’s significant advancements in natural language processing and visual understanding. The new interaction style and voice output make it feel more like a conversation with a human, while its coding capabilities and visual interpretation features make it a powerful tool for a wide range of applications.
The demonstration of ChatGPT 4o’s capabilities was further enhanced by the showcase of additional features, including real-time translations of conversations. This feature is made possible by the model’s availability in over 50 languages, as highlighted by Mira Murati. Another impressive feature displayed was the ability to read a person’s emotion based on a selfie image taken during the demo.
These new features will be accessible to all users, including those with free accounts, although with a limited number of prompts compared to pro or enterprise users. This move marks a significant expansion of ChatGPT 4o’s capabilities, making it a more versatile and user-friendly AI tool.
Mira Murati concluded the presentation by expressing gratitude to Janssen and Nvidia, the companies that provided the advanced GPUs necessary for the demo. Their contribution was instrumental in making the demonstration possible.
The rollout of ChatGPT 4o is expected to occur in the next few weeks, and users can look forward to additional announcements about the “next frontier” of GPT soon. This hints at further developments and innovations in the pipeline, making the future of AI technology even more exciting and promising.
Maybe you liked other interesting articles?