OpenAI launched the newest version of its ChatGPT artificial intelligence language model on Tuesday. GPT-4 has several new features. Including the ability to understand images as input and generate up to 25,000 words of text. Additionally, the model has access to a larger information database. Which enables it to provide more accurate information and write code in all major programming languages.
According to OpenAI, GPT-4 is much smarter than its previous version. As evidenced by its score on the Uniform Bar Exam, which was in the 90th percentile. Compared to its predecessor’s score in the 10th percentile. ChatGPT has already become the fastest-growing consumer application in history, with 100 million monthly active users in just a few months.
While OpenAI admits that ChatGPT is still “far from perfect,” the language model performs at a human level on various professional and academic benchmarks, such as scoring a 700/800 on the math SAT. However, AI language models have raised concerns regarding the accuracy of the information and the potential for bias.
OpenAI says that it has made ChatGPT safer by working with experts in AI safety and security to prevent disallowed content, such as illegal or morally objectionable content, from being generated by the model. GPT-4 is also 40% more likely to produce accurate information than its previous version, according to OpenAI.