OpenAI released the eagerly awaited update to ChatGPT – ChatGPT-4 – on Tuesday, with immediate access available on Microsoft’s Bing Chat and a developer waiting list.

ChatGPT-4 is “less capable than humans in many real-world scenarios”, OpenAI said in a recent blog post, but it represents “the latest milestone in OpenAI’s effort in scaling up deep learning”.

Since its launch last November, ChatGPT’s popularity has exploded, with traffic to the site hitting over one billion visits.

The new version of the tool is said to have “human-level performance on various professional and academic benchmarks”. ChatGPT-4 can pass the lawyer’s bar exam in the 90th percentile, for example, while by comparison, GPT-3.5’s score was in the bottom 10%.

Who Can Use ChatGPT-4?

Microsoft’s Bing Chat was co-developed with OpenAI and has run on GPT-4 for the last five weeks. GPT-4 is also available to OpenAI’s paying users, and developers can enter a waiting list to gain access to the platform.

What Are The Limitations Of ChatGPT-4?

“Despite its capabilities, GPT-4 has similar limitations as earlier GPT models,” OpenAI said, with the software still prone to producing bad advice, buggy code, or inaccurate information.

The company has cautioned against using the chat tool in “high-stakes contexts” and advises human review to check facts and add additional context. However, OpenAI claims a 40% improvement in error reduction.