OpenAI announces new and improved artificial intelligence models powering ChatGPT. (UCG/Getty Images).
- OpenAI has announced improvements to the model behind its chatbot ChatGPT.
- GPT-4o will now be able to integrate information from live video into its responses.
- OpenAI CEO Sam Altman said he felt the new model was “essentially different” from the original version of ChatGPT.
- For more financial news, visit: News24 Business top page.
OpenAI has unveiled a faster and cheaper version of the artificial intelligence model that powers its chatbot ChatGPT, as it looks to maintain its lead in an increasingly crowded market.
During a livestream event on Monday, OpenAI debuted GPT-4o. This is an updated version of his GPT-4 model from over a year ago. New large-scale language models are trained on vast amounts of data from the Internet, increasing their ability to process text, audio, and images in real time. Updates will be rolled out in the coming weeks.
When asked a verbal question, the system can respond with a voice response within milliseconds, allowing for more fluid conversations, the company said. In a demonstration of the model, OpenAI researcher and chief technology officer Miramurati conversed with his new ChatGPT using only voice, showing that the tool can return conversations. During the presentation, the chatbot translated audio from one language to another almost instantly, and at one point appeared to sing part of the story upon request.
“This is the first time we've made a huge leap forward in interaction and ease of use,” Murati told Bloomberg News. “We're enabling collaboration with tools like ChatGPT.”
This update provides free users with many features that were previously limited to users with paid subscriptions to ChatGPT. For example, the ability to search the web for answers to your queries, talk to a chatbot, and hear responses in a variety of voices. Run the command so that the chatbot saves the details so it can be called up in the future.
The release of GPT-4o will shake up the rapidly evolving AI landscape, where GPT-4 remains the gold standard. A growing number of startups and big tech companies, including Anthropic, Cohere, and Alphabet Inc.'s Google, have recently come out with AI models that match or exceed his GPT-4's performance on certain benchmarks. Masu.
@openai Introducing GPT-4o. GPT-4o is a new flagship model that can infer audio, vision, and text in real time. Check out our blog for more details.
? Original song – Aoi
The OpenAI announcement also came the day before the Google I/O developer conference. Google, an early leader in the field of artificial intelligence, is expected to use the event to announce further AI updates after aligning with Microsoft-backed OpenAI.
In a rare blog post on Monday, OpenAI CEO Sam Altman said that while the original version of ChatGPT offered hints at how humans use language to interact with computers, using GPT-4o “intuitive” He said that he felt it was completely different.
“It feels like an AI from a movie, but I'm still a little surprised that it's real,” he says. “Achieving human-level response times and expressiveness proved to be a game-changer.”
2x faster
Rather than relying on different AI models to process different inputs, GPT-4o (the “o” stands for omni) integrates voice, text, and vision into one model, making it more powerful than previous generations. It can be made faster. For example, if you enter an image prompt into the system, the system can respond with an image. The company says the new model is twice as fast and significantly more efficient.
“Having three different models working together introduces a lot of lag into the experience and breaks the immersion of the experience,” said Murati. “But having one model that natively infers audio, text, and vision cuts out all that latency and allows you to interact with ChatGPT the same way we interact with it today.”
However, the new model had some obstacles. As the researchers spoke during the demonstration, the audio frequently cut out. The AI system also surprised the audience by guiding the researchers through the process of solving an algebraic problem, then chimed in with a flirtatious-sounding voice that said, “Wow, you're wearing a really nice outfit.”
OpenAI today began rolling out new text and image features in GPT-4o to some paid ChatGPT Plus and Team users, and will soon bring those features to enterprise users. The company plans to make a new version of its “voice mode” assistant available to ChatGPT Plus users in the coming weeks.
OpenAI said that as part of the update, it will give everyone access to its GPT store, which includes customized chatbots created by users. Previously, it was only available to paying customers.
In recent weeks, speculation about OpenAI's next launch has become a Silicon Valley parlor game. A mysterious new chatbot has caused a stir among AI watchers after it appeared on a benchmarking website and appeared to match the performance of GPT-4. Altman mentioned X's chatbot with a wink, fueling rumors that his company was behind it. On Monday, an OpenAI employee confirmed on social platform X that the mysterious chatbot is indeed GPT-4o.
The company works on a wide range of products, including voice technology and video software. Bloomberg previously reported that OpenAI is also developing a search feature for ChatGPT.
On Friday, the company said it would not launch GPT-5 right away, putting some rumors to rest. GPT-5 is a long-awaited version of the company's model that some in the technology industry hope will be fundamentally more capable than current AI systems. He also said Monday's event will not announce any new search products or tools that could compete with Google. Google's stock price rose on the news.
But after the event, Altman quickly continued speculating. “I'll have more to share soon,” he wrote to X.