ChatGPT Vs. Google’s Gemini – Which Reigns Supreme?

This month, Google unveiled its latest attempt to dethrone ChatGPT from the position it’s held since it launched as king of the generative AI chatbots.

Bard – now renamed Gemini–was released in early 2023 following OpenAI’s groundbreaking LLM-powered chat interface. And to be honest, it’s often seemed as if it’s been playing catch-up.

Bard was capable of accessing the internet from day one thanks to its integration with Google’s search technology. Meanwhile, the launch version of ChatGPT was confined to the knowledge it was fed during their training.

But OpenAI soon added connectivity and the ability to access external information to ChatGPT via a hookup with Microsoft’s Bing. And connectivity aside, the consensus has always tended to be that ChatGPT is just more useful for a wider range of language processing tasks.

Now Google is pulling out the stops, rebranding Bard with the name of the language model that’s doing the work behind the scenes, and allowing access to its Advanced service via a subscription, priced to compete head-on with ChatGPT.

So, is it ready to step into the ring and go toe-to-toe with the undisputed champion? Here, I’ll give an overview of both platforms, highlighting the differences you’ll want to know about if you’re choosing which one to use.

The Language Models

First, it’s worth noting that both Gemini and ChatGPT are based on incredibly vast and powerful large language models (LLMs), far more advanced than anything publicly available in the past.

Remember, ChatGPT is just the interface through which users communicate with the language model – GPT4 (paying users of ChatGPT Pro) or GPT3.5 (free users.)

In Google’s case, the interface is called Gemini (previously Bard), and it’s used to communicate with the language model, which is a separate entity but is also called Gemini (or Gemini Ultra if you’re paying for the Gemini Advanced service).

Something important to take into consideration is that although we call them both chatbots, the intended user experience is slightly different. ChatGPT is designed to enable conversations and help solve problems in a conversational manner – much like chatting with an expert on a subject.

Gemini, on the other hand, seems designed to process information and automate tasks in a way that saves the user time and effort.

From a technical perspective, the power of LLM models is often measured by the number of parameters (trainable values) within the neural network. It’s been reported that GPT-4’s networks contain around a trillion parameters, but no solid facts are known about the number of parameters used by Gemini.

This might not be important, however, as it may be enough to just know that both are very, very powerful.

AI professor at Arizona State University, Subbarao Kambhampati, recently told Wired, “We have basically come to a point where most LLMs are indistinguishable on qualitative metrics.”

In other words, the technical size and power of the model isn’t what’s important – it’s how it has been tuned, trained and presented to help users solve problems that really matters.

And The Winner Is …

After using both for a while to hold various conversations on different topics, it seems clear to me that ChatGPT is still the more powerful chat interface, thanks to the grunt provided by GPT-4. Gemini is closing the gap, though!

Information Retrieval

One advantage of Gemini is that by default, it considers all of the information at its fingertips – including the internet, Google’s vast knowledge graph, and its training data.

ChatGPT, on the other hand, will often still choose to try and answer a question solely relying on its training data. This can occasionally lead to out-of-date information. However, you can circumvent this by prompting it to search the web to get the latest and most up-to-date data. But this is still introducing an extra step that Gemini has shown is not really needed.

In my experience of using both platforms, I would have to say that Gemini proves to be slightly more adept than ChatGPT when it comes to online searching and integrating the information it finds into its responses.

When ChatGPT does head online and look for information, its responses tend to lose some of their dynamism. It often seems as if it will answer questions or provide responses based on a single web search and a single source of information rather than conducting a comprehensive analysis of all the information it can access and coming to a conclusion.

Here’s a quick example of what this means. I often use AI chatbots to give me a quick overview of a company or its products or services. Using the same prompt (“tell me about [URL]”), ChatGPT will often simply regurgitate a marketing blurb from the website.

In the brief time I’ve had to test it, Gemini seems to take a more nuanced approach. It summarizes the information it can find while attempting to generate a balanced overview of features.

So, I would say that this is one area where Gemini edges slightly ahead of its rival.

But that’s far from the end of the story. When it comes to intelligently parsing the information it’s been trained on in order to formulate a response, ChatGPT still comes out as the winner.

And The Winner Is…

Let’s call this one a draw, with Gemini being better when it comes to formulating answers from online text and ChatGPT being better at no-internet queries.

Multi-Modal Capabilities

Multi-modal AIs are those that are capable of processing more than one type of data. Early versions of ChatGPT only read and generated text. But since OpenAI upgraded its “engine” to GPT-4, it gained the ability to process visual and audio data, making it multi-modal. Gemini, on the other hand, was multi-modal out of the box (although not all of its features were immediately activated).

ChatGPT generates images using the DALL-E model, which was also developed by OpenAI. Gemini, on the other hand, utilizes Google’s Imagen 2 engine. Both are clearly very powerful and can generate amazing results. However, I would say that ChatGPT is more consistent when it comes to creating an image that closely matches what I was looking for when we compare them on a same-prompt basis.

One difference that’s been noted by others is that Imagen 2 and Gemini are slightly better at producing photorealistic, very highly detailed images. ChatGPT, on the other hand, excels when it comes to managing spatial relationships between objects in its images, and it is better at creatively interpreting prompts.

Both are also capable of understanding and writing computer code across a huge range of programming languages. There are slight differences in how they do this, though.

Now, I am not a programmer – but the great thing is, with ChatGPT or Gemini in front of you, you don’t need to be.

There’s no doubt that ChatGPT’s superior conversational abilities give it some significant advantages here. If you aren’t quite sure what your code should do or about the best way to integrate it, it’s better when it comes to generating clear and helpful guidance and offering suggestions and tips.

And The Winner Is …

I’m going to give this one to ChatGPT again. While Gemini does create better photorealistic, ChatGPT wins when it comes to generating images that closely match what the user is asking for with their prompt. Gemini seems slightly better at creating technical code but can’t match ChatGPT as a conversational interface to use while building and experimenting.

(Just a quick note: Gemini image generation hasn’t yet launched for users in Europe – hopefully, it will be added soon.)

So Which Is Best?

Well, neither is by any means perfect. Both still suffer from hallucinations and will, fairly frequently, provide information that is simply wrong. For example, Gemini told me that OpenAI’s Dall-E 2 doesn’t use diffusion model technology (it does.) And ChatGPT told me that Gemini isn’t capable of generating images (it is).

But for my money, if you’re only going to subscribe to one, I’d be inclined to go for ChatGPT Pro at the moment.

There are a few caveats – if you’re heavily into Google’s ecosystem, then Gemini’s ability to interface with Gmail and Google Docs is likely to be a star attraction for you. Similarly, if you’re an experienced coder and your main need is coding, definitely check out Gemini (but also take a look at Microsoft’s Co-Pilot).

For writing and creating documents, summarizing, general-purpose image generation and learning through conversations, I’d say ChatGPT is better right now. For this reason, it retains its place as the best that’s currently available.