Majority of Copilot users prefer GPT-3.5's speed over GPT-4's performance, reveals Microsoft

Reading time icon 1 min. read


Readers help support MSpoweruser. We may get a commission if you buy through our links. Tooltip Icon

Read our disclosure page to find out how can you help MSPoweruser sustain the editorial team Read more

In a recent tweet, Mikhail Parakhin, the President of Search and Advertising at Microsoft, shared insights into Copilot usage patterns, revealing that the majority, i.e., around 70% of users, opt for the Balanced mode, the fastest and default option, which utilizes GPT-3.5 instead of the more advanced GPT-4. The other modes – Precise and Creative mode, which uses GPT-4 – get 15% of users each.

Parakhin attributes this trend to the general tendency among users to prioritize speed and convenience over exploring alternative options, even when they offer significant performance advantages, which GPT-4 does over GPT-3.5.

When it comes to text generation, different situations call for different approaches. For common tasks such as searching, a Balanced approach works best – it’s fast and efficient. However, when a user needs to generate new content that’s more creative, expressive, and longer, a Creative approach is more suitable, even if it takes a bit longer. The best approach is to use a Precise method for factual writing that requires precision and minimal conjecture.

I underestimated the popularity of Precise mode over Creative mode.

What mode do you usually prefer?

More about the topics: copilot