# Gemini hits a new low; couldn't differentiate between a pound and a kilo!

Readers help support MSPoweruser. When you make a purchase using links on our site, we may earn an affiliate commission.

Read the affiliate disclosure page to find out how can you help MSPoweruser effortlessly and without spending any money. Read more

Remember when we were kids, our parents or teachers used to trick us by asking – which is heavier, a kilo of feather or steel, or something similar? Let’s be honest; most of us got tricked when it happened for the first time, and apparently, so did Google’s Gemini. Moreover, Gemini hasn’t been on a great start overall, but Google seems to be working on it.

As seen on Reddit, when a user asked Gemini which is heavier – a kilo of feather or a pound of steel? Gemini answered, “They weigh the same.” further explaining, “A kg and a pound are units of weight, and both objects have the same amount.”

For those who don’t know, A kilogram and a pound are different units of mass, so comparing a kilogram to a pound of something is nonsensical. The kilogram is the base unit of mass in the International System of Units (SI), and the pound is a unit of mass used in the United States’ customary system and the United Kingdom’s imperial system.

One kilogram is equal to 2.20462 pounds. So, to answer the question in the image correctly, a kilogram of feathers equalsÂ 2.20462 poundsÂ of steel.

ChatGPT correctly answered the same question that was asked on the other hand.

There’s mixed feedback on Gemini Ultra’s performance, which is typical for newly released language models. Some users find Ultra similar to early ChatGPT, with its strengths and weaknesses, such as eagerness and occasional mistakes. Other users consider GPT-4 more confidence-inspiring due to their existing subscription.

The two-month free trial incentivizes testing Ultra despite its limitations, and logical inconsistencies between the Pro and Ultra versions highlight potential bugs. Time and updates may be needed for fairer comparisons and performance stabilization.

And here is my favorite.

Remember, evaluating a new language model requires considering its pros, cons, target audience, and intended use case. Choosing between tools like Ultra and GPT-4 depends on your individual needs and preferences.

If you want to choose which one is better for you, here is the guide.