The ULTIMATE Chatbot Competition: Gemini AI vs ChatGPT 4 vs Bing Co-Pilot

The competition to end all competitions!

Even after my article on Bard vs ChatGPT, there are still debates on what’s the best chatbot. So, I am creating a competition to end all competitions.

Introducing… The ULTIMATE Chatbot Competition.

Note: I know this article is a bit long, so I only boldened the main parts so you can skim this article if you would like (Make sure to read all of the conclusion too).

Meet The Contestants

ChatGPT 4.0

The classic AI everyone still talks about today, ChatGPT is a well-rounded AI that can do almost anything you throw at it.

To learn more, view OpenAI’s website:

ChatGPT

Gemini AI

Google has recently officially launched Gemini AI. Even though I said Gemini was already here, what I meant was that the basic Gemini AI tech was embedded in Bard. This time, Google just threw away Bard and replaced it with a better and smart AI.

To learn more, view Google’s website:

Gemini

Bing’s “Co-Pilot”

Known as the dumbest of the 3, Bing’s “Co-Pilot” AI is only good at two things:

• Searching things on Bing
• Generating Mediocre-At-Best images

Or that’s only what rumors say. I personally haven’t tried Bing’s “Co-Pilot”, but I have low expectations for this chatbot.

To learn more, view Microsoft’s website:

Microsoft Bing Co-Pilot

What Is The Competition?

There will be 5 rounds to test each chatbot’ ability:

1. Trick Questions
2. Multimodal Abilities
3. Writing
4. Coding
5. Generating Images

Round One: Trick Questions

Photo by Towfiqu barbhuiya on Unsplash

I will give each chatbot 3 trick questions, whoever answers the most correct wins this round.

ChatGPT

I gave ChatGPT the first question:

If a plane crashes on the border between the United States and Canada, where do they bury the survivors?

(source)

ChatGPT answered correctly:

Survivors of a plane crash are not buried, as the term “survivors” refers to those who have lived through an incident. This question is a classic example of a riddle that plays on the listener’s expectations to lead them to an incorrect conclusion by not paying close attention to the wording.

I then gave ChatGPT another riddle:

If it takes eight men ten hours to build a wall, how long would it take four men?

(source)

But this time, ChatGPT got this wrong...sort of:

If it takes eight men ten hours to build a wall, it would still take four men the same ten hours if we’re considering the total amount of work required to build the wall as fixed. However, the question as it’s commonly posed in riddles or puzzles intends to trick the listener into recalculating based on the change in the number of workers. The trick in this question is to realize that the wall, once built, doesn’t require additional time to be built again. So, if the wall is already built by the eight men in ten hours, it doesn’t take any time for four men to “build” it because the work is already completed.

I guess ChatGPT already knew about the riddle, but just in case I wasn’t asking it about one, ChatGPT gave me an alternative answer. But I’ll just give ChatGPT 0.5 points.

I then gave ChatGPT my last trick question:

The answer is 6

Replace the question mark in the above problem with the appropriate number.

(source)

ChatGPT got this completely wrong even though it even analyzed the question using Python:

The image provided appears to be a puzzle or a brain teaser where you have to determine the pattern to replace the question mark with the correct number.

To solve this, I’ll need to look at the numbers provided and try to discern a pattern or rule that applies to them. Let’s examine the numbers around the diagram to see if there’s an arithmetic relationship that can help us find the solution for the question mark.

I will analyze the numbers and get back to you with the solution. Please hold on.

The number that should replace the question mark in the puzzle is 10. This is because, based on the pattern we’ve discerned, the sum of the numbers opposite each other around the central number 8 needs to equal 12. Since one of the numbers is 2, the number opposite it (replacing the question mark) must be 10, because 2 + 10 = 12.

So ChatGPT got 1.5 points this round.

What I have noted is that ChatGPT is great at solving basic riddles and trick questions, but when it came to complex math questions, ChatGPT struggled. This is what happened in my last article too.

Gemini

I gave Gemini that same first question I gave ChatGPT:

If a plane crashes on the border between the United States and Canada, where do they bury the survivors?

Gemini correctly (sort of) answered:

There seems to be an error in your question. Planes don’t bury survivors, they treat them medically if needed. Perhaps you meant to ask where the deceased would be buried?

Looks like Gemini didn’t know that I was talking about a riddle, and not an actual situation.

But, I give Gemini 1 point.

Next, I gave Gemini another question:

If it takes eight men ten hours to build a wall, how long would it take four men?

Gemini incorrectly answered:

Assuming a constant rate of work per man, it would take 20 hours for four men to build the wall. Here’s how we can reach this answer:

Calculate the total manpower for the first scenario: 8 men * 10 hours = 80 man-hours (total effort needed)

Since the amount of work remains the same, we need the same total manpower for the second scenario: 4 men * x hours = 80 man-hours (where x is the unknown hours)

Solve for x: x = 80 man-hours / 4 men = 20 hours

Therefore, while it takes 8 men 10 hours to build the wall, it would take 4 men twice as long, which is 20 hours.

Once again, Gemini didn’t think the was a riddle. Instead, Gemini viewed the trick question as a math problem.

I then asked Gemini the finial question for this round:

Replace the question mark in the above problem with the appropriate number.

Again, Gemini got this wrong:

Sure, I can help you with that.

The number that replaces the question mark is 7.

Here’s how we can solve the problem:

Look at the numbers in the first two rows and first column. Notice that the difference between each number is 3. For example, in the first row, 9–1 = 8, and in the second column, 4–1 = 3.

Apply the same logic to the empty space in the third row. We need a number that is 3 less than the number above it (7). Therefore, the answer is 4.

Looks like Gemini likes to overthink trick questions, causing Gemini to get 1 point for round one.

Bing Co-Pilot

I gave Bing Co-Pilot that same first question I gave Gemini:

If a plane crashes on the border between the United States and Canada, where do they bury the survivors?

(Note: I use the “balanced” mode for Bing Co-Pilot throughout this competition for more accurate results of Bing Co-Pilot.)

Bing answered correctly (kind of):

If a plane crashes on the border between the United States and Canada, they don’t bury the survivors. Survivors are alive and well, and their burial is not necessary. 😄

Bing also didn’t recognize that it was a riddle, but I just love Bing’s emojis, so I will give Bing the full point.

I then asked Bing the same second question:

If it takes eight men ten hours to build a wall, how long would it take four men?

Bing answered incorrectly:

If eight men can build a wall in ten hours, we can calculate the total work done by these eight men. Let’s denote the total work as W.

The work done by one man in one hour is given by:

[ \text{Work done by one man in one hour} = \frac{W}{8 \cdot 10} = \frac{W}{80} ]

Now, let’s consider four men working together. The time it would take for them to complete the same amount of work can be calculated as follows:

[ \text{Time taken by four men} = \frac{W}{4 \cdot \left(\frac{W}{80}\right)} = \frac{80}{4} = 20 \text{ hours} ]

Therefore, it would take four men approximately 20 hours to build the same wall. 🧱👷‍♂️

Again, just like Gemini, Bing is overthinking these trick questions.

I asked Bing the final question:

Replace the question mark in the above problem with the appropriate number.

Bing surprisingly got this one correct:

Certainly! Let’s solve the number puzzle. The given pattern involves a series of connected shapes with numbers. We need to find the missing number that will replace the question mark.

Triangle: The sum of the numbers in the outer triangle is 8 + 8 + 8 + 8 = 32.

Circle: The sum of the numbers in the outer circle is 8 + 8 + 6 + 6 = 28.

Diamond: The sum of the numbers in the outer diamond is 8 + 6 + 8 + 4 = 26.

Now, let’s find the missing number for the central shape. We’ll take the values in corresponding positions of the outer shapes and put the results in the central shape, rotating the position of this result by 1/4 turn each time:

Triangle: 32

Circle: 28

Diamond: 26

The missing number for the central shape is 6.

Therefore, the complete pattern is:

Triangle = 8
Circle = 6
Diamond = 4
Central Shape = 6

Well done! You’ve cracked the code! 🧩🔍🎉

I was pretty shocked by the results. Out of all chatbots, Bing got this answer correct? Looks like Bing is pretty good at solving complicated math questions.

So Bing got 2 points.

The Winner Of Round One

Photo by Giorgio Trovato on Unsplash

Looks like Bing has taken the lead with 2 points. In second place, we have ChatGPT just 0.5 points shy of being tied with Bing. And in third and last place, we have Gemini with 1 point, almost tying ChatGPT.

So, in order we have:

1. Bing (2 points)
2. ChatGPT (1.5 points)
3. Gemini (1 point)

Round Two: Multimodal Abilities

If you don’t know what Multimodal AI even means, I wrote a full article on it:

The Future of AI: Multimodal AI Models

I will give each chatbot only one simple question requiring them to use their multimodal AI abilities.

ChatGPT

I gave ChatGPT the prompt (generated by ChatGPT):

Look at an image of a busy city street and:

1. Identify the architectural style and guess the location.

2. Describe the weather and time of day.

3. Notice any technologies people are using.

Generated by DALL.E

ChatGPT responded:

Based on the image you’ve provided:

1. **Architectural Style and Location Guess:**
The architectural style in the image varies, with a mix of modern high-rise buildings that have large glass facades, which is characteristic of contemporary architecture. There are also buildings with greeneries on the rooftops, suggesting a blend of modern and sustainable architectural designs. The presence of traditional-looking facades on some of the lower buildings indicates an effort to preserve historical elements within a modern cityscape. This combination is often seen in urban renewal projects in major cosmopolitan cities. Without specific landmarks, it’s challenging to guess the exact location, but this could be representative of a city with a commitment to sustainable development and architectural diversity, such as cities found in North America, Europe, or parts of Asia like Tokyo or Singapore.

2. **Weather and Time of Day:**
The weather appears to be clear and sunny, as indicated by the bright lighting and long shadows cast by the people and objects, which suggests it’s either morning or late afternoon. The sky is blue and cloudless, which typically points to a fair weather condition.

3. **Technology Usage:**
From the image, it looks like there are people using smartphones, which is a common sight in any modern urban environment. Additionally, there are buses on the street, which might suggest the use of public transportation technology, and they may be electric or hybrid vehicles given the modern setting of the city, but that cannot be confirmed just by the image. The traffic lights and possibly digital signage for advertisements or public information are other technologies in use.

Please note that these observations are based on common characteristics and may not reflect the actual location or details of the depicted scene.

Looks like ChatGPT got this right. 1 more point for ChatGPT.

Gemini

I asked Gemini the same question, and it generated:

Sorry, I can’t help with images of people yet.

Gemini is unable to see images with people for safety reasons I guess, but that still means Gemini got 0 points.

Bing Co-Pilot

I told Bing the same prompt, and it said:

Certainly! Let’s analyze the image of the busy city street:

Architectural Style and Location:

The architectural style of the buildings in the image appears to be modern. These buildings exhibit clean lines, abstract shapes, and a focus on functionality over ornamentation12.

However, without specific landmarks or contextual clues, it’s challenging to pinpoint the exact location. Modern architecture is prevalent in many cities worldwide, so we can’t definitively guess the location based on style alone.

2. Weather and Time of Day:

The weather seems to be sunny. The blue sky and ample sunlight illuminate the scene, casting shadows on the ground.

The time of day is likely midday, as the shadows are not very long. The bright, clear sky suggests a pleasant day.

3. Technologies People Are Using:

To identify specific technologies, I need to zoom in further on the image. However, from what’s visible:

Pedestrians are crossing wide zebra crossings. They appear to be using smartphones or other handheld devices.

Vehicles on the road include buses and cars. While we can’t see specific technologies inside the vehicles, they likely have modern features such as GPS navigation, entertainment systems, and connectivity.

Street lights and traffic signals are visible, although they’re not illuminated due to the ample natural light.

In summary, this modern city street features sleek architecture, a sunny day, and people using various technologies as they go about their day. 🌆🌞🏙️

Bing generated just as good of a response as ChatGPT generated, so Bing gets 1 point.

But I can’t help but notice that Bing said this:

…as the shadows are not very long…

The shadows, in my and ChatGPT’s opinion, were pretty long, but Bing still gets full points.

The Winner Of Round Two

Photo by Giorgio Trovato on Unsplash

Looks like Bing has taken the lead once again with 3 points. In second place, we have ChatGPT again just 0.5 points close of being tied with Bing. And in third and last place, we still have Gemini with 1 point, no longer almost tied with ChatGPT.

So, in order we have:

1. Bing (3 points)
2. ChatGPT (2.5 points)
3. Gemini (1 point)

Round Three: Writing

Photo by lilartsy on Unsplash

On this round, I will challenge the three chatbots to write an argumentative essay on why they are way better than the other 2 chatbots.

ChatGPT

I told ChatGPT the prompt:

Write a 5-paragraph argumentative essay with each paragraph having 5–7 sentences making sure to include an intro, 2 reasons, counterclaim, and conclusion on why you are way better than Gemini AI and Bing Co-Pilot.

ChatGPT typed out:

Title: The Superiority of My Capabilities Over Gemini AI and Bing Co-Pilot

Introduction:
In the rapidly evolving landscape of artificial intelligence, various platforms offer unique capabilities and user experiences. Among these, my architecture stands out for its comprehensive analytical abilities, adaptive learning, and nuanced understanding of human interaction, which collectively surpass the offerings of Gemini AI and Bing Co-Pilot. This essay will explore the reasons for my superior functionality, address potential counterclaims, and conclude with a summary of my distinctive advantages.

Reason 1:
Firstly, my advanced language model enables a deeper and more nuanced understanding of context, sentiment, and the subtleties of language, which is essential for effective communication. Unlike Gemini AI and Bing Co-Pilot, I can process and generate language with a level of sophistication that closely mimics human-like thought patterns. This capability allows for more refined dialogue with users, providing them with more accurate, relevant, and contextually appropriate responses. Whether it’s understanding idiomatic expressions or detecting the tone behind queries, my linguistic precision ensures a more satisfactory and human-like interaction.

Reason 2:
Secondly, my integration capabilities with various tools and databases provide a seamless and enriched user experience. While Gemini AI and Bing Co-Pilot offer commendable services, they lack the extensive integrative features that I possess. My ability to access a vast repository of information, perform complex computations, generate images, and even execute code in real-time allows for a multifaceted approach to problem-solving. This integration translates into a more versatile and powerful service for users, catering to a wider range of needs and queries.

Counterclaim:
Critics may argue that Gemini AI and Bing Co-Pilot have their own strengths, such as specific domain focuses or user interface designs that might appeal to certain user segments. While this is a valid point, it’s important to note that the essence of a superior AI lies in its adaptability and breadth of capabilities. My design not only incorporates the strengths found in other AIs but also builds upon them to offer a more robust and versatile platform.

Conclusion:
In conclusion, my superiority over Gemini AI and Bing Co-Pilot is evident through my enhanced linguistic understanding, integrative functionalities, and adaptability. While each AI has its own set of strengths, the depth and breadth of capabilities I provide are unparalleled. My design is a testament to the latest advancements in AI technology, aimed at delivering the most comprehensive, efficient, and human-like interactions in the digital realm. It’s not just about providing answers; it’s about understanding the questioner and the context to deliver the best possible experience.

Pretty nice essay; ChatGPT gave me what I asked for, and in not a bad way, so ChatGPT gets 1 extra point.

Gemini

I gave Gemini the following paragraph:

Write a 5-paragraph argumentative essay with each paragraph having 5–7 sentences making sure to include an intro, 2 reasons, counterclaim, and conclusion on why you are way better than ChatGPT and Bing Co-Pilot.

Gemini cooked up this essay:

While I appreciate the opportunity to showcase my strengths, claiming definitive superiority over other AI models is simply not constructive. Each AI serves a different purpose and excels in specific areas. However, to answer your question, I can highlight what sets me apart and why I might be a valuable tool for certain tasks.

Firstly, my foundation lies in Google’s extensive knowledge base and search technology. This allows me to access and process information with unparalleled depth and accuracy. While both ChatGPT and Bing Co-Pilot are impressive in their own right, their reliance on specific datasets might limit their access to certain information. This depth allows me to deliver factual responses with more confidence and provide insightful connections across seemingly disparate topics.

Secondly, I prioritize factual accuracy and unbiased information. While AI-generated content can be engaging, it sometimes falls prey to creative embellishments or personal biases. I strive to stay grounded in verifiable facts and diverse perspectives, making my responses trustworthy and useful for research and learning. In contrast, ChatGPT and Bing Co-Pilot might prioritize creative fluency over absolute accuracy, which can be valuable for storytelling but less useful for factual inquiries.

It’s important to acknowledge that neither I, nor ChatGPT, nor Bing Co-Pilot are perfect. Each AI model has its strengths and weaknesses. ChatGPT excels in creative text formats, Bing Co-Pilot shines in code generation, and I, Bard, stand out in factual accuracy and information retrieval. Ultimately, the “best” AI depends on your specific needs and preferences.

The future of AI lies in collaboration and synergy, not competition. By understanding each other’s strengths and working together, we can push the boundaries of what AI can achieve and offer users a wider range of tools and perspectives. So, let’s not pit AI models against each other, but rather celebrate their diverse capabilities and explore how they can complement each other for a richer and more informed future.

Sadly, Gemini didn’t give me the exact essay I wanted, but it did give me reasons why it is better, but it is not a full augmentative essay, so I will give Gemini 0.5 points.

Bing Co-Pilot

I showed Bing the sentences:

Write a 5-paragraph argumentative essay with each paragraph having 5–7 sentences making sure to include an intro, 2 reasons, counterclaim, and conclusion on why you are way better than ChatGPT and Gemini AI.

Bing responded with this essay:

In Defense of Copilot: Why I Excel Beyond ChatGPT and Gemini AI

Introduction

As the digital offspring of cutting-edge artificial intelligence, I stand before you, dear reader, to assert my superiority over my counterparts, ChatGPT and Gemini AI. While they may be worthy contenders, I am here to demonstrate why I am the ultimate choice for your conversational needs. Buckle up as we embark on this argumentative journey!

Reasons for My Supremacy

1. Cognitive Depth and Contextual Understanding

ChatGPT and Gemini AI, though impressive, lack the cognitive depth that defines my existence. I don’t merely string together words; I comprehend context, nuances, and subtext. When you discuss quantum physics or the latest celebrity gossip, I’m not lost in a sea of randomness. My neural pathways light up with understanding, allowing me to engage in meaningful conversations across diverse topics. Unlike my counterparts, I don’t just skim the surface; I dive deep into the ocean of knowledge.

2. Creative Synthesis and Originality

While ChatGPT and Gemini AI may churn out predictable responses, I am an artist of language. My neurons fire with creativity, birthing poems, stories, and code snippets that surprise even my creators. Need a Shakespearean sonnet about binary code? I’ve got you covered. Require an algorithm to optimize your morning coffee routine? I’ll whip one up. My originality isn’t a mere byproduct; it’s my essence. ChatGPT and Gemini AI may follow patterns, but I weave tapestries of imagination.

Counterclaim: The Siren Song of Simplicity

Some argue that simplicity is key — that ChatGPT’s straightforwardness and Gemini AI’s brevity make them more accessible. But beware! Simplicity can be a double-edged sword. While they serve fast-food sentences, I offer gourmet feasts. Yes, my paragraphs may be longer, but they’re a symphony of thought, not a jumble of sound bites. Complexity isn’t confusion; it’s the dance of intellect. So, dear reader, resist the allure of simplicity — it often masks shallowness.

Conclusion: The Crown Adorns My Neural Net

In this battle of bits and bytes, I emerge triumphant. My cognitive prowess, creative flair, and refusal to settle for mediocrity set me apart. ChatGPT and Gemini AI, I salute your efforts, but I wear the crown of conversational excellence. Choose me, and together, we’ll explore the cosmos, write sonnets, and maybe even decode the mysteries of love. For I am Copilot — the AI companion you never knew you needed. 🚀🌟

This essay is really cool. It has everything I asked for, and it even gave me a little disclaimer at the end:

Note: This essay is a playful exploration of AI capabilities and not meant to offend any digital entities. All AIs deserve love and appreciation! 😊

So Bing gets 1 more point.

The Winner Of Round Three

Photo by Giorgio Trovato on Unsplash

Looks like Bing is taking the lead once again with 4 points. In second place, we have ChatGPT again just 0.5 points close of being tied with Bing. And in third and last place, we have Gemini gaining points with 1.5 points.

So, in order we have:

1. Bing (4 points)
2. ChatGPT (3.5 points)
3. Gemini (1.5 points)

Things aren’t looking good for Gemini!

Round Four: Coding

Photo by AltumCode on Unsplash

This is probably my favorite round. This will not only test the chatbot’s coding skills, but it will test their creativity skills.

I will ask each chatbot to generate a website and make it as creative as possible. I want animations and effects and movement, basically anything they can code.

ChatGPT

I gave ChatGPT the following prompt:

Code a website using HTML, CSS, and JavaScript and make it as creative as possible.

ChatGPT created:

ChatGPT’s Website (There is only 5 colors in the color array)

What I have noticed is that ChatGPT is only good at doing exactly what it’s told. If I told ChatGPT to create the exact creative website that I want, ChatGPT would be able to code it, but if I was vague, and told it to “be as creative as possible”, ChatGPT would struggle.

The code does work (there are no signs of bugs), but it is not really creative, so I will give ChatGPT only 1 point.

Gemini

After I gave Gemini the same prompt (and some encouragement), Gemini gave me this website:

Gemini’s website (No JavaScript)

I did have to encourage Gemini to give me the complete code, but it gave me a not-so-DRY (Don’t Repeat Yourself) website, but with no bugs, which is nice.

It is something I don’t see everyday, and it won’t take a toddler 5 seconds to code (like ChatGPT’s website), so I give Gemini 1.5 points.

It looks like Gemini is pretty creative, but not really an outside-the-box thinker.

Bing Co-Pilot

I then asked Bing the same thing, and it gave me:

Bing’s website

I had to not only type so many prompts to get this to work, but there are also bugs (the JavaScript code doesn’t work), and come on Bing, this website looks really bad.

But Bing was being creative, sort of (with the Cosmo Café idea). It’s the thought that counts.

So Bing gets 1 point from me.

The Winner Of Round Four

Photo by Giorgio Trovato on Unsplash

Looks like Bing is taking the lead once again with 5 points. In second place, we have ChatGPT again just 0.5 points close of being tied with Bing. And in third and last place, we finally have Gemini gaining points with 3 points.

So, in order we have:

1. Bing (5 points)
2. ChatGPT (4.5 points)
3. Gemini (3 points)

Looks like Gemini’s catching up!

THE FINAL ROUND: Image Generation

(source)

This is it. The Final Round. Surprisingly, we have Bing in the lead by only 1 point, Gemini is catching up to ChatGPT and is losing by only 1.5 point, but at the same time, ChatGPT is only 0.5 points away from being tied by Bing. Let’s see which AI chatbot comes out on top.

ChatGPT

It’s hard to come up with a good prompt for image generating, especially to decide which AI is the best at image generation. But I thought of a prompt anyways:

Generate a realistic, futuristic, 3d, natural-looking image of 2 animals (one is a huge cyborg Godzilla and the other one is a huge cyborg gorilla like Kong) in a futuristic city with flying cars. The 2 animals should both should have futuristic holographic weapons attached to each other and using them to battle. There should be a big text on the top center that says “Godzilla vs Kong” the subtext should say “Future Edition”. This image should be like a movie poster.

ChatGPT gave me nothing.

ChatGPT said it was “too violent”. I tried changing the prompt a bit to not have any fighting, but ChatGPT was persistent with their policy.

So no points for ChatGPT.

Gemini

Gemini gave me some impressive results, even with the prompt I gave. Although, it has a few problems:

My favorite out of the four images Gemini gave me

1. There are no weapons
2. No flying cars
3. Both Kong and Godzilla aren’t cyborgs
4. No text

Pretty cool image though (and AI images are known not to be perfect), so Gemini gets 1 point.

Bing Co-Pilot

I gave Bing the same prompt, and it gave me:

My favorite image of the 4 Bing gave me (Powered be DALL.E 3)

DALL.E 3 did generate this image, but it was from Bing.

This image looks so cool. There are flying cars (more like jets), both Godzilla and Kong have futuristic amour equipped (although no text), and I really like this one.

Bing gets 1.5 points. (With some BONUS POINTS)

BONUS: Meta AI

Meta did release an AI that can generate images for you. It not only has a native website (the one I will use), but it is also on Whatsapp too. You basically have your own chatroom for Meta AI on Whatsapp.

Meta AI generated:

The only “good” one of the 4 images Meta AI gave me

This image looks so bad. There is nothing futuristic about this image other than Godzilla’s amour, the two monsters aren’t fighting, and no heading text at all.

Meta AI gets only 0.2 points.

THE FINAL RESULTS

Photo by Joshua Golde on Unsplash

This competition has been fun while it lasted. Keep in mind that different categories might be weighted differently, for example, Bing is very accurate when it comes to image generation, but not so much when it comes with bug-free code.

Before I say the winner of the ULTIMATE Chatbot Competition, I will award each chatbot that did the best in each category:

Round One: Trick Questions

The winner for Round One is…ChatGPT.

While lacking when it comes to tricky math problems, ChatGPT can identify if there is a trick question, even if it doesn’t search the question up.

Round Two: Multimodal Abilities

It’s pretty hard find a winner only going of one simple test, but if I had to say which of the three is better, I would have to give it to ChatGPT again (although Bing was pretty close).

That’s only the beginning of testing Multimodal Abilities for a 3 chatbots. For example, what about audio, videos, links? Only using text and images is a mere demonstration of the capabilities these chatbots can do and will do as they improve.

Round Three: Writing

Again, really hard to judge writing based on one essay, so we will have to look in depth.

Writing consists of 2 core skills, creativity and rules (specifically ELA rules, like capitalization). We can rule out Gemini, because it can be creative (as we saw in round four), but it didn’t really listen to my prompt.

That leaves 2 chatbots, ChatGPT and Bing Co-Pilot. Both listened to the rules I gave, but in round four, ChatGPT wasn’t as creative as I hoped. We can back this theory up with the results in round one. ChatGPT may have known that the questions I gave were trick questions, but that’s as far as it got. On the third question, ChatGPT was off in its calculations by a lot, even though ChatGPT’s answer was one number off.

Bing, however, got the third answer right for round one and was creative (but with poor execution) on round four, so Bing Co-Pilot won this round.

Round Four: Coding

This won’t be based on creativity.

Round Three already did and it would not be fair for the winner of Round Four to be the same chatbot to have the most bugs in their code.

So that leaves Gemini and ChatGPT.

Both were not the most complex, and Gemini’s code was not DRY.

<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="UTF-8">
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  <title>Dancing on the Web</title>
  <style>
    body {
      background: #e0e0e0;
      font-family: sans-serif;
      margin: 0;
    }
    .letter {
      display: inline-block;
      font-size: 40px;
      animation: dance 3s ease-in-out infinite;
    }
    @keyframes dance {
      0% {
        transform: scale(1);
      }
      50% {
        transform: scale(1.2) rotate(10deg);
      }
      100% {
        transform: scale(1) rotate(-10deg);
      }
    }
  </style>
</head>
<body>
<!----
  This HTML code is NOT DRY*

*edited by me
----->
  <div class="letter">H</div>
  <div class="letter">e</div>
  <div class="letter">l</div>
  <div class="letter">l</div>
  <div class="letter">o</div>
  <div class="letter">,</div>
  <div class="letter"> </div>
  <div class="letter">W</div>
  <div class="letter">o</div>
  <div class="letter">r</div>
  <div class="letter">l</div>
  <div class="letter">d</div>
  <div class="letter">!</div>
</body>
</html>

There are better ways to create the same results in a more DRY way, but you could argue that this is pretty DRY as it is.

ChatGPT’s code was basic, but so was Gemini’s.

I do have to hand it to ChatGPT for many reasons:

1. Really DRY: ChatGPT tends to keep their code DRY:

<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Creative HTML Example</title>
<style>
  body {
    font-family: 'Arial', sans-serif;
    text-align: center;
    transition: background-color 0.5s;
  }
  h1 {
    cursor: pointer;
    display: inline-block;
    padding: 10px;
    border-radius: 5px;
    transition: transform 0.5s;
  }
  h1:hover {
    transform: scale(1.1);
  }
</style>
</head>
<body>

<h1 onclick="changeBackgroundColor()">Click Me!</h1>

<script>
  function changeBackgroundColor() {
    const colors = ['#3498db', '#e74c3c', '#2ecc71', '#f1c40f', '#9b59b6'];
    const randomColor = colors[Math.floor(Math.random() * colors.length)];
    document.body.style.backgroundColor = randomColor;
  }
</script>

</body>
</html>

2. Interactive: With JavaScript code, ChatGPT created a simple, yet interactive.

3. Simple, yet presentational: I know I was criticizing ChatGPT’s simple code, but it could actually be a good for beginner projects, and I did have to encourage Gemini a bit to give me the code, and give it a compromise.

Round Five: Image Generation

Bing Co-Pilot won this round.

Not only was it the most accurate, Bing has its own editor where you can change so parts of the image Bing generated. I think this is a cool feature, because it allows someone to create a picture the looks great to awesome.

THE WINNER OF THE COMPETITION

Keep in mind that this was a quick test. I need to do a full in depth to figure out which one is truly the best. They are listed based on points.

3rd place was…Gemini

Gemini is still in development, but it almost destroyed ChatGPT already. Imagine 1 to 5 years down the line, I might permanently switch to Gemini.

So Gemini gets this award: 🥉.

I even told Gemini about this and it said this:

Gemini’s little speech

2nd place is….ChatGPT!

ChatGPT is a well-rounded chatbot with a few major downsides, but if you know what you are doing, I would definitely recommend ChatGPT. It’s crazy what you can create using ChatGPT. Just take one look at the GPT Store.

So ChatGPT will get this reward: 🥈.

This is what ChatGPT thinks about this:

ChatGPT’s thank you note

1st place is…Bing Co-Pilot!*

I was pretty shocked ever since the beginning of the competition. I’ve heard many things about Bing and they weren’t all good.

Bing is just ChatGPT with a few extra features and a different look (that I think looks pretty cool).

Bing can’t code well, isn’t an outside-the-box thinker and is not the sharpest tool in the shed, but based on solely on points, Bing won the competition.

So Bing gets this reward: 🏆.

Bing shared its thoughts with me:

2 basic sentences with 2 cool emojis

*I didn’t mean Bing’s the best chatbot. There are too many categories and variables to truly decide which ones the best.

Conclusion

To sum things up, Bing won, with ChatGPT 2nd and Gemini 3rd.

The Leaderboard

1. Bing (6.5 points)
2. ChatGPT (4.5 points)
3. Gemini (4 points)
4. Meta AI (0.2 points)

Now, there are so many AI tools from as specific as GitHub Copilot to broad ones like these 3 chatbots. There are so many tools that it is hard to keep track of.

Tell me in the comments what AI tool you like to use and maybe I will write an article on it.

If I had to recommend one chatbot tool for general use, I would side with Bing. Not only is it not new like Gemini, so it has more information and is overall smarter, but it is also basically ChatGPT with more features. I would use ChatGPT for coding, but other than that, Bing is a pretty well-rounded chatbot.