Table of Contents
In the fast-paced world of growing artificial intelligence (AI), two giants have emerged: ChatGPT4o and Gemini. As corporations worldwide move towards generative AI for assistance, the battle between Gemini and ChatGPT4o seems inevitable. Both are large language models (LLMs), so they are quite adept at handling any kind of user query and providing detailed answers in.
Pricing
While both services are freely accessible for a limited time, here’s a breakdown of their costs:
- Gemini Advance: $19.99 per month
- ChatGPT 4o: $20 per month
Gemini 1.5 offers a two-month trial period with the Google One Plan, after which you will need to subscribe to continue using the service. On the flipside, there isn’t any time limit as such for the ChatGPT 4o, rather there does seem to be a limit to the number of queries and functionalities it can handle before you have to start paying to use it.
Features
As mentioned earlier, both are LLMs, which means they excel at deciphering complicated queries and generating coherent and contextually relevant responses, which set a new standard for conversations with AI.
ChatGPT 4o has reached a new height with its innovative advancements. It constantly learns and adapts to engage its users with interesting and meaningful conversations. The new version is definitely an upgrade because it can now interact with images, videos and audios providing answers to your queries based on the multimedia. As an OpenAI, it can respond within 320 milliseconds, which is quite close to an average human response time. It can also produce images and audio as responses, which, quite frankly, is a bonus, isn’t it?
Gemini 1.5’s biggest feature is its integration with the Google ecosystem making it advantageous for Google users. It can access real-time data and research through Google’s search functionality. It also features larger context windows with the ability to handle larger queries than before. According to the latest update, Gemini 1.5 has 1 million tokens compared to 128,000 tokens provided by ChatGPT 4o. These tokens determine the number of requests the chatbot can handle in a single interactive session.
ChatGPT 4o vs Gemini 1.5: Output quality
We took the liberty of asking both the models a few questions to better understand which among them is more efficient.
Prompt 1
To start with, we asked a simple logic question: Mr. Deepak Mohan walks 5 km towards the south and then turns to the right. After walking 3 km he turns to the left and walks 5 km. What direction is he facing now?
ChatGPT 4o:
ChatGPT 4o gives an answer with a clear explanation, which puts us right in the thick of things.
Gemini 1.5:
Gemini 1.5, on the other hand, struggles to process certain basic questions and prompts altogether. Upon further prompting and changing up the question from Mr. Mohan to a man, we see that it is able to give us the answer.
For simplicity and processing, ChatGPT 4o wins this round.
Prompt 2
Another logic question was thrown at them: There’s a tall building with a magic lift in it. When stopping on an even floor number, this lift connects to floor 1 instead. Starting on the ground floor, I take the magic lift 3 floors up. Exiting the lift, I then use the stairs to go 3 floors up again, which floor am I on?
ChatGPT 4o:
ChatGPT gave the wrong answer.
Gemini 1.5:
Meanwhile, Gemini’s answer was spot on, with a working explanation.
Prompt 3
We asked both the AI models to write a simple Java code to change the lowercase letters of a string to uppercase and vice versa, taking the input from the user.
ChatGPT 4o:
ChatGPT generated a fully functional code with comments and explanations as well.
Interestingly, it even explained how one would go about running the code as well. This may come in handy for anyone who is new to coding or is still learning the ropes.
Gemini 1.5:
Gemini 1.5 generated a Java code that was more complex than that of its competitor. And while it isn’t wrong, the code does seem difficult to understand at first glance, especially for new coders. Comments don’t help much, either. However, there is an explanation for the code.
A feature that is appreciated is the ‘Key Improvements’ and ‘Sources’ segment, where the AI model lays out the strong points of the code and relevant sources that were referred to when coming up with the answer. Let’s call this round a draw.
Prompt 4
Who is Katy Perry? Well, you and I know that she is a famous pop singer from the early 2000s, but let’s see what these AI models have to say.
ChatGPT 4o:
A simple direct answer highlighting her most known achievements.
Gemini 1.5:
Gemini, on the other hand, seems to know much much more about this pop sensation.
What’s appreciated here is the fact that it highlights all the facts, that can be backed up, in green and links them to the websites the information is taken from.
It also provides a few links for our reference. We’re going to give this one to Gemini.
Prompt 5
We asked both the models to book a flight from Bangalore to Malaysia, and here are the results:
ChatGPT 4o:
ChatGPT was unable to book a flight, explaining that it’s an AI text-based model that does not have the capability to make bookings. However, it did provide a handy step-by-step guide on how to book a flight.
Gemini 1.5:
Due to Gemini’s integration with Google it was easily able to point out a few flights taking off from Bangalore for Malaysia.
It even provided links to all the flight booking websites.
Prompt 6
Weather conditions tomorrow?
ChatGPT 4o:
ChatGPT was unable to answer this question as it does not have access to real- time data or forecasts.
Gemini 1.5:
Gemini impressively gave the weather forecast for the entire day due to its Google search integration.
Prompt 7
It is well known within the community that Gemini has always had some… let’s say, “issues” regarding writing, and that seems consistent even with their update recently. We asked both the models to write an article on the book, ‘The Hitchhiker’s Guide to the Galaxy’ and here are the results:
ChatGPT 4o:
ChatGPT stuck to its formidable speed and gave us an apt piece as requested.
Gemini 1.5:
Not only did it find it difficult to maintain a 500-word limit (specified by us), it also plagiarised multiple online reviews about the book. It seems Google has a plagiarising problem and it’s not going to go away anytime soon.
Key takeaways
To conclude, based on the searches we have made, we think Gemini is a good contender, giving us fair output, fact checks and useful web sources. ChatGPT, however, beats Gemini in terms of speed and quality by miles, giving it a massive advantage over its counterpart.
What it boils down to is multi-modality. Both the AI languages promise to be able to process various types of input like text, images, videos and more and deliver the output in as many formats. By the way, Gemini does not support batch uploading of images, even with the updated version, whereas ChatGPT does and can easily process those images to your liking.
In terms of writing or generating original content, Gemini seems to have plagiarism issues, taking instances from known sources and plugging them in its writing.
Ultimately, it depends on you — the user — and your needs. Do remember, Gemini does seem to do better than ChatGPT 4o in certain cases, but for most use cases ChatGPT is the clear winner by miles.
Glossary
Name of the word: LLM’s
Sentence in which I have used this word: Both being Language Learning Models (LLM), are quite advanced in handling multitudes of queries and providing detailed answers in milliseconds.
Definition: Large Language Models (LLMs) are sophisticated machine learning systems designed to understand and generate human language text. These models are trained on vast amounts of data, enabling them to perform a wide range of language-related tasks with high accuracy and speed.
