ChatGPT, the artificial intelligence platform of the OpenAI company, has been a hot topic in recent months, and now the company has come up with a new product. According to the company, its latest technology is “multimodal”, meaning that both images and text requests can stimulate content generation. The text input feature will be available to ChatGPT Plus subscribers and software developers with a waiting list, while the image input capability is for research purposes only for now.
OpenAI’s latest technology is in some cases a huge improvement over the previous version known as GPT-3.5, the company said. According to OpenAI, the new model performed in the top 10% of test takers during a simulation of the bar exam required for US law school graduates to practice law before starting to practice law, while the older model scored in the bottom 10%.
While the two versions may seem similar in casual conversation, “the difference emerges when the complexity of the task reaches an appropriate threshold,” OpenAI said, noting that “GPT-4 can handle more reliable, creative, and much more nuanced instructions.”
GPT-4 is 82 percent less likely to respond to requests for prohibited content than its predecessor and scores 40 percent higher on certain factual tests, the company said.
Cover photo: Jakub Porzycki/NurPhoto via Getty Images
Leave a Comment