Google Gemini vs ChatGPT: Which AI model is better?

With Google Gemini finally replacing Bard it is exciting to see if it is better or worse than ChatGPT. Can Google finally overtake OpenAI or is it just another inferior version of ChatGPT?

Although keep in mind as Gemini has only just been released it is too early to say if it is any better, but testing it myself we will try to compare them to the best of my ability.

What is Google Gemini?

Google Gemini is a large language model family developed by Google DeepMind, succeeding previous models like LaMDA and PaLM 2. It was announced in December 2023 and is considered a competitor to OpenAI’s GPT-4.

It’s a powerful AI model capable of understanding and generating different types of information, including text, code, audio, images, and video. This “multimodal” ability makes it versatile for various tasks.

What is ChatGPT?

ChatGPT is an artificial intelligence language model developed by OpenAI. It’s based on the GPT (Generative Pre-trained Transformer) architecture, which uses deep learning techniques to understand and generate human-like text. 

OpenAI’s ChatGPT is designed to perform a wide range of language tasks, such as answering questions, providing explanations, engaging in conversation, and generating creative content.

Google Gemini Pros

• Multimodality: Gemini’s core strength lies in its ability to understand and integrate multiple data types. It effortlessly combines text, images, audio, video, and code, offering comprehensive insights and solutions.
• Performance: Gemini Ultra has outperformed human experts in MMLU (Massive Multitask Language Understanding), excelling in 30 of 32 academic benchmarks. It shows greater capabilities in areas like natural image and audio understanding, and mathematical reasoning.
• Advanced Coding: Gemini is adept at understanding and generating high-quality code in popular programming languages. It demonstrates exceptional performance in coding benchmarks, including HumanEval and Natural2Code.
• Scalability and Efficiency: Designed for Google’s Tensor Processing Units (TPUs), Gemini operates efficiently across platforms, from data centers to mobile devices (Gemini Nano). Its scalability makes it suitable for a wide range of applications and user needs.
• The output seems more human: When comparing creative text, Google Gemini seems to do a better job of producing text that doesn’t feel entirely AI-made.

Again, a disclaimer: Goggle Gemini has only just been launched, so it is too early to tell how good it is. It will take at least a few weeks to see how it compares to ChatGPT. So take this comparison with a pinch of salt.

Google Gemini Cons

• Image generation: Compared to Dall-E it is quite worse and compared to Midjourney it is even worse. It is okay for now, but nothing special.
• Factual errors even after a double check: Even after double-checking Gemini’s information with its function, it may still provide style errors or misunderstand the information.
• No opt-out for training: You cannot choose to not be included in data used for future training of Gemini, raising questions about user privacy and control.

Google Gemini Pricing

Gemini Basic Plan:

• Price: Free
• Features: Includes the Pro 1.0 model, suitable for everyday needs. Assistance with writing, planning, learning, and more. Integration with familiar Google apps.

Gemini Advanced Plan:

• Price: $20 per month (2 month free trial available).
• Features: Gemini Advanced is available as part of Google One AI Premium Plan and includes: Ultra 1.0 model, Google’s most capable AI model. Designed for the best performance and highly complex tasks. Future integration with Gemini in Gmail, Docs, and other Google services. Additional benefits include 2 TB of storage and other Google One benefits. 

ChatGPT Pros

• Problem-Solving: GPT-4’s expanded knowledge base and problem-solving abilities allow it to tackle complex queries with higher accuracy.
• Creative Collaboration: It excels in creative tasks, such as songwriting, screenplay drafting, and adapting to individual writing styles.
• Visual Input Processing: Capable of understanding and responding to images, GPT-4 broadens the scope of interaction through image-based inputs.
• Extended Text Handling: With the ability to process over 25,000 words, it supports long-form content creation and detailed document analysis.
• Superior Reasoning: GPT-4 surpasses previous models in reasoning, demonstrating higher performance in tasks requiring nuanced understanding.
• Safety and Alignment Improvements: Through extensive training and feedback integration, GPT-4 shows a significant reduction in producing disallowed content and improved factual accuracy.
• Continuous Learning: Incorporating real-world usage feedback, GPT-4 evolves continuously for enhanced performance and safety.

ChatGPT Cons

• Inaccuracy: ChatGPT can sometimes fabricate facts or provide factually incorrect information. This can be due to several factors
• Hallucinations: Even when factually wrong, ChatGPT can sometimes present its answers with high confidence, making it difficult for users to discern truth from fiction. The simplest example of this is to ask ChatGPT how many Bs are in Banana. ChatGPT will say 3 for some reason.

• Dall-E: It is quite good at the moment. But like many other generators that create images, it cannot provide correct text or images, people have weird characteristics (Midjourney is much better) and it requires quite along text prompts to get the images you want.
• Text output: Every text ChatGPT provided feels a lot like ChatGPT made it. It is quite difficult to describe, but it uses words that no human would ever use when writing a sentence.

ChatGPT Pricing

• Free plan available
• Plus Plan: $20 per month, offering access to GPT-4, image and voice chat capabilities, and custom GPT creation.

Gemini vs ChatGPT: What is the difference between Gemini and ChatGPT?

Well, the price is almost identical, the main difference I would say is ChatGPT’s image generator is miles ahead but Google’s Gemini provides much better creative text.

The benchmarks have their uses, but saying something is a few percentage points better doesn’t really say much for real-world use.

Gemini Ultra vs GPT-4

As we all know GPT-4 has been released for almost a year now, so it has had quite a lot of improvements to make it better since it was released.

Although Gemini Ultra has just been released, according to benchmarks it is better and faster than GPT-4. I agree that even the standard and ultra versions produce results much faster than GPT-4, but time and use will tell if it is better.

Gemini vs ChatGPT: Is Google Gemini better than ChatGPT?

I had some hopes for Gemini but only expected it to provide some sort of alternative to GPT-4. It’s not equivalent, but if you’ve only ever used GPT-3.5, Ultra is certainly an improvement.

Personally, I think Gemini’s AI chatbot is better for creative writing. I asked it to write several different fictional stories and the level of quality is higher than GPT. More unpredictable, stronger stories.

GPT has a very noticeable writing style that she defaults to, which is easy to notice. This exaggerated style and overuse of certain phrases such as “delves”. 

But it’s been known for a while that Google has been beaten. How quickly we remember Google’s panic and developers’ “all is lost” mood when GPT-3.5 was released, and the desperate rush to catch up that followed. Complete with them rushing Bard to market, only for GPT to come out a few days later and crush it and everything else.

In the long run, we will just have to wait and see how Gemini improves. The short answer in my opinion is that it is worse than ChatGPT.

Gemini vs ChatGPT: Community Concerns

Here are some of the most interesting takes on Reddit I found about Gemini vs ChatGPT:

• “I have been playing around with it and comparing it to GPT4, and all across the board, GPT4 is much more accurate, seems to have a much greater knowledge base, and does not hallucinate as much.”
• “They scooped a lot of the best minds for this specific problem and have been dedicated to working on it longer than anyone else at this scale. Its not easy for anyone to catch up to OpenAI”
• “So far it’s fine, if a little underwhelming. The lack of message limits is nice.”

I think it is important to see other people’s reactions and not just read about Gemini on Google’s website or from YouTube videos of people who have tried it for a short time. Don’t take those comments too seriously, try it for yourself.

Final thoughts

I enjoy using both AI models extensively, they have their strengths and weaknesses. Although in the end ChatGPT still comes out on top. 

Gemini is better for some specific tasks like content creation, but that is only a small part of what an AI chatbot can do. However, one thing that stands out is that with Bard you can check where the information is coming from and with ChatGPT it is a bit more difficult.

I would advise you to try both and see which one you like better.