Tuesday 14 March saw OpenAI begin the release of the newest iteration of its artificial intelligence (AI) model – GPT-4. The new model is “multinodal”, accepting images as prompts to generate content alongside text input which its predecessor, GPT-3, was limited to.
Sam Altman, CEO of OpenAI, revealed the new LLM on Twitter, calling it OpenAI’s “most capable and aligned model yet”. Altman also noted that GPT-4 is “still flawed, still limited”, in a warning that there is still a long way to go for the development.