GPT-4 is not just a text generator. It is also multimodal, which means it can understand and produce different types of media, such as images, music and video. For instance, you can ask GPT-4 to translate a text into another language with an accompanying image or audio clip.
GPT-4 is the successor of GPT-3, which was released in 2020 and was already considered a breakthrough in AI. GPT-3 had 175 billion parameters (the numbers that define how the model works),
while GPT-4 has more than 1 trillion parameters. This makes GPT-4 more powerful and capable of learning from more data.