GPT-4 outperforms, outclasses and outscores 90% of humans in world’s toughest exams

GPT-4 outperforms, outclasses and outscores 90% of humans in world’s toughest exams

Mehul Reuben Das March 15, 2023, 13:03:16 IST

During its showcase of GPT-4, OpenAI revealed that they had a modded version of ChatGPT running on GPT-4 take some of the toughest exams in the US. GPT-4 performed better than most humans on those exams, outscoring about 90% exam takers.

read more
Advertisement
GPT-4 outperforms, outclasses and outscores 90% of humans in world’s toughest exams

GPT-4 has been making all sorts of headlines in the hours since it has been released, mainly because of the unbelievable abilities that the new AI language model has. When OpenAI’s Sam Altman unveiled the GPT-4, they dubbed the program as “more reliable, creative, and able to handle much more nuanced instructions” than its predecessor, GPT-3.5. During the launch, OpenAI showcased a number of demos that showed what GPT-4 is actually capable of, and how it is better than the previous generation of ChatGPT, which was based on GPT3.5.

In one such demo, the engineers at OpenAI revealed that a modded version of ChatGPT that was running on GPT-4, had outscored and outperformed and outscored 90 per cent of humans in some of the toughest exams in the US. “GPT-4 is a large multimodal model (accepting image and text inputs, emitting text outputs) that, while less capable than humans in many real-world scenarios, exhibits human-level performance on various professional and academic benchmarks,” it said.   “For example, it passes a simulated bar exam with a score around the top 10 per cent of test takers; in contrast, GPT-3.5’s score was around the bottom 10 per cent,” said OpenAI’s representatives. Also read: Here are 5 ways the new GPT-4 absolutely spanks the current version of ChatGPT The AI model also performed at the 93rd percentile on a SAT reading exam and at the 89th percentile on an SAT math test, the company added. Furthermore, OpenAI’s data shows that GPT-4 cleared LSAT with 88 per cent, GRE quantitative at 80 per cent and GRE Verbal with a staggering 99 per cent. GPT-4, basically, passes almost all the exams that it took, that too with flying colours.   Because of how well the AI bot has performed in these demos, several startups have signed up to use the API of GPT-4 in their apps.   Language-learning program Duolingo, which is aiming to create conversational bots, as well as automated tutors for the online education company Khan Academy are just some of the educational companies that have signed up to use OpenAI’s new GPT-4 in their products. “AI accelerates our mission to make high-quality education available to everyone in the world. The things that we can do now with the power of OpenAI’s technology are going to shape the future of education,” said Luis von Ahn, CEO and Co-founder of Duolingo.   “Most people don’t have access to a one-on-one human tutor, but I believe AI will allow us to eventually recreate the experience of a human tutor and scale it to everyone in the world. I’m so excited to share these incredible new features with millions of Duolingo learners,” the CEO added. Read all the  Latest News Trending News Cricket News Bollywood News, India News and  Entertainment News here. Follow us on  FacebookTwitter and  Instagram.

Latest News
Find us on YouTube
Subscribe

Top Shows

Vantage Firstpost America Firstpost Africa First Sports