Californian start-up OpenAI has presented a new version of the technology behind its ChatGPT chatbot.
Dubbed GPT-4, the new generation generates better results than previous versions, the company said in a blog entry late on Tuesday.
"GPT-4 is a large multimodal model (accepting image and text inputs, emitting text outputs) that, while less capable than humans in many real-world scenarios, exhibits human-level performance on various professional and academic benchmarks," OpenAI said.
Paying customers will be given access to GPT-4 but there is already a wait list, according to the company.
Last year's launch of OpenAI's ChatGPT - a generative AI chatbot developed with support from Microsoft that answers prompts with polished essays, poetry, computer code and other human-like content - has thrust artificial intelligence into the public spotlight.
Users simply type a request into a chat box and the system can generate a response almost instantly.
For the GPT technologies, the software captured enormous amounts of text and images. On this basis, it can formulate sentences that are hardly distinguishable from human conversations. In doing so, the programme estimates which words might follow next in a sentence. Among other things, this basic principle carries the risk that the software "hallucinates" facts, according to OpenAI.
While it has been significantly reduced compared to previous models, this limitation still exists with GPT-4, the company says in the blog entry.
"GPT-4 generally lacks knowledge of events that have occurred after the vast majority of its data cuts off (September 2021), and does not learn from its experience," OpenAI said.
Meanwhile the new version is also said to be good at analyzing and describing images, a feature still under review.
