After months of rumors, OpenAI has announced GPT-4: the latest in a series of AI language models that power apps like ChatGPT and the new Bing.
The company claims that the model is "more creative and collaborative than ever" and that it "can solve difficult problems with greater precision, thanks to its broader general knowledge and problem-solving abilities."
OpenAI says it has already worked with several companies to integrate GPT-4 into their products, including Duolingo, Stripe and Khan Academy. The new model will also be available in ChatGPT Plus as an API.
A research post on the OpenAI blog states that the distinction between GPT-4 and its predecessor GPT-3.5 is "subtle" in casual chat (GPT-3.5 is the model that powers ChatGPT), but that the differences between the two systems are clear when used in more complex tasks.
The company says these improvements will be seen in the GPT-4's performance on a number of tests and benchmarks, including the Uniform Bar Exam, LSAT, SAT Math, and SAT Evidence-Based Reading & Writing.
In the tests mentioned GPT-4 scored as you can to see here. Rumors of GPT-4 and its capabilities have been rife over the past year, with many stating that it would be a huge step up from previous systems.
"People are begging to be disappointed and they will be disappointed," said OpenAI CEO Sam Altman, in an interview in January.