When Microsoft announced the new AI-enabled Bing, which it built on top of OpenAI’s GPT models, neither company would confirm which version of GPT was being used beyond saying it was a next-gen version of the model that powered ChatGPT. Today, OpenAI announced GPT-4, a significant update to GPT-3.5. As it turns out, Bing was using it all along.
“We are happy to confirm that the new Bing is running on GPT-4, which we’ve customized for search,” Microsoft’s Yusuf Mehdi, the company’s corporate VP and consumer chief marketing officer, wrote in today’s announcement. “If you’ve used the new Bing preview at any time in the last five weeks, you’ve already experienced an early version of this powerful model.”
When Microsoft launched the new Bing, there were plenty of rumors that it ran GPT-4 already, so this isn’t a huge surprise, but it is interesting to see that Microsoft was confident enough in the model to stake its reputation on it — and to pay the bills for this more complex model. Microsoft, it’s worth noting, is using a combination of GPT-4 and its own Prometheus model in order to provide more up to date information and put guardrails around OpenAI’s model.
After a rougher start than the company probably anticipated (in part because the new Bing was prone to hallucinations), Microsoft quickly iterated on the new Bing in recent weeks and after placing a number of restrictions on it early on, the company is now loosening up again. Only yesterday, Microsoft extended the number of possible turns in a conversation to 15 and now allows users up to 150 chats per day.
So if you want to give the new GPT-4 model a try, just head over to Bing (or get on the waitlist, if you haven’t done so yet).
Bing itself, by the way, still insists that it doesn’t run on GPT-4. Somebody needs to tell Sydney that it’s not under NDA anymore…