What is GPT-4?

Computers have long struggled to work with ordinary human language: reading it, summarizing it, translating it, and writing it. Large language models are one answer, machine learning systems trained on a vast amount of text so they can generate, summarize, translate and analyze text in many settings. GPT-4 is a large language model from OpenAI. It can take both images and text as input and produces text as output, so a person can ask questions or give instructions in plain language, optionally alongside a picture, and get a written reply. OpenAI published it on March 14, 2023, describing it as the latest milestone in its effort to scale up deep learning (training very large neural networks).

Written from primary sources and reference pages, and checked against them by an independent AI fact-checker. Drafted with AI assistance.