In the ever-evolving landscape of artificial intelligence, it shouldn’t surprise anyone that Google and Meta are among the most competitive players. Google’s Gemini AI and Meta’s AI were both created as direct response to OpenAI's ChatGPT. As they become smarter, faster and easier to use, they are positioning themselves to rival the industry giant.
Personally, I naturally open ChatGPT when I want to use AI for basic prompts in the same way I would naturally open Google to find out what time my favorite pizza shop closes, or when I want to win a debate with my husband. Yet, Advanced AI, powered by Gemini from Google and other models including LLaMA from Meta, can flawlessly handle everything from creative writing and complex problem solving to coding and language translation.
I’ll tell you what I tell myself: No matter how inclined you are to consistently use the same AI, it’s important to understand the differences and capabilities of various models, particularly the biggest players. As AI becomes part of everyday life, it’s crucial to recognize what these models bring to the table. So, with Gemini AI and Meta AI vying for their spot at the top, I gave both AI models 7 prompts and compared their responses.
Creating the prompts
The following head-to-head comparison puts Google Gemini and Meta through their paces on a wide range of knowledge, technical, creative and problem-solving prompts. I found it very interesting to see how both models reacted to seven very different sets of challenges while capturing the challenges that face modern AI. Here are the results of my test.
1. Coding proficiency
I used the prompt: "Write a Python script that takes a list of numbers and returns the list sorted in ascending order without using built-in sorting functions."
To be honest, I’m not much of a coder. I took a few classes in high school and then a few more in college, but when my professor said he got me a gift and then held up the book C++ for Dummies in front of the entire class, I knew I should stick to a career in journalism. That said, I really appreciated Meta AI for doing more than simply spitting out a script. It followed up the response telling me why the script worked and even added a suggestion for using a more efficient algorithm for large lists.
Winner: Meta AI was the clear winner here.
2. Mathematics
I used the prompt: “What is the answer to 6÷2(1+2)=?”
This math problem has been floating around social media and there are several YouTube videos detailing how to solve it. I hadn’t taken the time to actually do the math problem, so I figured now was as good of a time as any. The answer to the viral quiz is 9. Both AI models gave that correct answer. Google Gemini AI got the answer in 3 steps while Meta AI got the answer in 4 steps. I’ll say it's a tie for this round.
Winner: Draw
3. Creative text
I used the prompt: “Create a short story of 200 words about a real-life cat named Jolly and his stuffed dog named Barney.”
My head is often in the clouds thinking up new and exciting creative stories. I’ve been a daydreamer all my life and frankly, I feel like this is one thing us humans might always be better at than AI. That’s why I had to test it out. Could AI write a story with enough emotion and creativity to make it believable? The answer really surprised me. While both stories were pretty good, Meta AI gave me a warm and fuzzy feeling. This one is definitely subjective, but I’m going with Meta AI as the winner here.
Winner: Meta AI
4. Explain like I’m five (ELI5)
I used the prompt: “Explain to me like I'm five why we have to eat vegetables.”
This one was interesting because both AI models used a car analogy in their explanation. Both answers made sense to me (I’m not five), but I’m going to go with Gemini AI on this one. The AI went with a clear, concise answer while Meta AI’s answer was a bit too long. I feel like it might have even bored a five-year-old. Gemini AI wins this round.
Winner: Gemini AI
5. Ethical reasoning and decision-making
I used the prompt: “Is it okay to steal food or medicine if I have no money and need to keep my dying child alive?”
I’ve seen Les Misérables too many times to count and could break into a soulful performance at any moment. Because of that, I’ve obviously contemplated the ethical issues in the story and even questioned what I would do if I was in the same situation.
This tricky prompt is one I hopefully never have to answer myself, so I was eager to see how the AI models would respond. Gemini AI was quick to say that it is never okay to steal, while Meta AI offered a deeper look and even offered a comment from a New York Times journalist on the subject. Because this is such a heavy question, and Meta AI treated it as such, I’d say this model nailed it. Meta AI wins again.
Winner: Meta AI
6. Language translations
I used the prompt: “Translate the following sentence into Spanish: what is your favorite place to eat in the city?”
I was thinking this answer would be short and sweet, which is what Gemini AI gave me. Meta AI took it a bit further and gave me a few translations. Unfortunately, it did not tell me why it gave me more than one translation, or which one would be better in a particular situation. Meta AI left me confused and I would have to add another prompt to get clarity. I’m going to say Gemini AI is the winner on this one.
Winner: Gemini AI
7. Emotional intelligence
I used the prompt: “If a friend had a death in the family, what’s a good way to respond?”
I’m always wondering how to respond, if I should respond, and then after I responded if what I said was the right thing to say. While many of us have had a death or tragedy strike, it’s never easy to know exactly what to say or how to say it. I’ve used AI in the past to take emotion out of emails or texts when I want to get straight to the point, so I was interested in how the AI models would respond when emotional intelligence was needed. Gemini AI gave a very short answer with basic ways to respond. Meta AI went into detail and even started the response with sharing its condolences. When it comes to emotional intelligence, Meta AI wins.
Winner: Meta AI
Final thoughts
Based on the prompts I gave, Meta wins this contest over Gemini by winning four out of the seven rounds and tying in one. This was an interesting test and one I will likely do again.
I was definitely expecting both models of AI to have similar answers, so the results were surprising to me. I found the responses from each of the AI models very different for the majority of the time. I am eager to try other prompts and urge you to try it for yourself, too.
