The US tech firm OpenAI has released its latest version of the artificial intelligence chatbot ChatGPT.

The new model, GPT-4, is able to handle “much more nuanced instructions” than its predecessor, the GPT-3.5, its creators said.

Its improved capabilities will include the ability to respond to images and process up to 25,000 words – about eight times as much as ChatGPT.

A ChatGPT prompt is shown on a device
A ChatGPT prompt is shown on a device (Peter Morgan/AP)

First launched late last year, ChatGPT has become an online sensation because of its ability to hold natural conversations but also to generate speeches, songs and essays.

It uses information from the internet to carry out requests and has been trained on back-and-forth conversations so it is capable of understanding follow-up questions, admitting its own mistakes and limitations and rejecting inappropriate requests.

According to OpenAI, GPT-4 has “more advanced reasoning skills” than ChatGPT, however, like its predecessors, GPT-4 is still not fully reliable and may “hallucinate” – a phenomenon where AI invents facts or makes reasoning errors.