Gemini vs ChatGPT: How Does Google’s Latest AI Compare?

Google has just taken the next step in its AI journey with the debut of Gemini 1.0, its latest and greatest AI model.

Can Google’s best and brightest AI beat its most well-known competitor, the generative AI chatbot ChatGPT? Here’s what to know about the newest entry in the battle of the bots.

In this guide, you’ll learn how Gemini and ChatGPT compare when it comes to their history, availability, pricing, and functionality. 

What Is Gemini AI?

You can be forgiven for getting confused about how the biggest corporations are using AI. Both Google and Microsoft have such long histories as corporate business software behemoths that the total number of product lunches, rebrands, and acquisitions at each is now virtually impossible to count.

This means that AI tools must bolster and supplement a huge range of wildly differing products and services. So, when Google refers to “Gemini,” it can mean a lot of different things.

Google has released three different tiers of the Gemini large language model (LLM) that all constitute “Gemini 1.0,” the first iteration of the software. These are:

• Gemini Ultra — Most powerful (but slower) model
• Gemini Pro — Most scalable, all-purpose model; currently used for Bard
• Gemini Nano — Most efficient model. Less powerful, but great for on-device tasks

Currently, Google has released Gemini Pro for use in Bard in tandem with the current AI model that Bard uses, which is (confusingly) also called Bard. Google has also released Gemini Nano for the Pixel 8. The Pixel 8 Pro is the first phone designed to work with Nano, which makes new features possible, including a “Summarize in the Recorder” app and Smart Reply for WhatsApp.

Gemini Ultra is not out yet: Google is still fine-tuning it, with plans to release it for a new version of Bard, Bard Advanced, sometime in 2024.

Google will roll out far more uses of Gemini across 2024 as well. They haven’t offered up a full list yet, but have mentioned that it will cover four major product areas: Search, Ads, Chrome, and Duet AI. Google says it is already “experimenting” with Gemini in Search, using it to reduce latency by 40%.

Gemini vs ChatGPT: Which Is Better?

Google claims Gemini can beat ChatGPT. In fact, outperforming ChatGPT is a huge selling point. But does the current Gemini-powered Bard live up to the hype?

Well, here’s a catch: It’s really only the “Ultra” version of Gemini that Google claims can beat ChatGPT. Early reports confirm that Ultra does beat ChatGPT-4 narrowly in many categories.

However, that version won’t be ready until some time in 2024. Right now, people are comparing Gemini Pro with ChatGPT’s free version, GPT-3.5, and they don’t seem to be thrilled with the results.

The Verge’s Nilay Patel posted that he “asked the new Gemini-powered Google Bard an important question and it confidently hallucinated a Vergecast interview that doesn’t exist complete with a link.” Writing for The Atlantic, Matteo Wong says that Gemini can beat GPT-4 “on most metrics,” but that it’s just an “iterative advance” — it’s a tiny improvement, not a giant leap forward.

Granted, ChatGPT is far from perfect. We’ve discussed its penchant for AI hallucinations in the past, and we have pointed out exploits that must be fixed as well. Still, Google has to prove that Gemini can actually impress ChatGPT users. So far, it has not.

Gemini vs ChatGPT Pricing

Gemini is a value-add for Google. It’s going to be folded into a mind-boggling number of products, from the Search engine everyone uses to the Chrome browser that everyone uses Search with. Many of those iterations are free to use — although you’re technically paying for the privilege with your ad data.

Currently, you can use Gemini in some limited ways totally for free. The generative text AI platform Bard is the easiest way to familarize yourself with the tool, but you can also use Gemini with Pexel 8. But the AI tool will be included in paid services in the future. We have yet to see if it will be a paid add-on, or which enterprise services will feature it.

Meanwhile, ChatGPT has an even easier pricing system. Anyone who signs up for a free account can get the free tier by sharing their email address and phone number. A paid tier is also available to all, ChatGPT Plus. It costs $20 a month, and gets you a more advanced LLM plus the ability to integrate ChatGPT plugins (some of which will cost an additional fee). Finally, there’s the Enterprise ChatGPT tier, with on-demand pricing.

Google currently uses versions of AI in free products that are used by billions of people everyday: Product including Search, YouTube, Gmail, Google Maps, Google Play, and Android. In the not-so-far-off future, Google is probably quietly roll out Gemini across these services. On one hand, you won’t have to pay extra for a monthly subscription. On the other hand, chances are good that you won’t be able to escape using Gemini.

Gemini vs ChatGPT: Data Sources and Language Models

The Bard chatbot platform is now powered both by the Bard LLM and the Gemini Pro LLM. The full rollout of the Gemini family across all Google products will be a slow affair, and will likely involve a similar combination of previous LLMs with the newer, more efficient Gemini LLMs. For the average user, it’s all happening under the hood.

When Ultra debuts in 2024, it is poised to take the mantle from ChatGPT: It has already outperformed the OpenAI model across 30 out of 32 categories, tests have shown.

“Our most-capable model, Gemini Ultra, advances the state of the art in 30 of 32 benchmarks, including 10 of 12 popular text and reasoning benchmarks, 9 of 9 image understanding benchmarks, 6 of 6 video understanding benchmarks, and 5 of 5 speech recognition and speech translation benchmarks.” –Jeff Dean, Chief Scientist, Google DeepMind

ChatGPT’s data model comes from publically available data, scraped straight from the internet. Both the GPT-3.5 and GPT-4 models use data from September 2021 or earlier.

Your Generative AI Explainer Guide

Here at Tech.co, we’ve been reviewing and comparing all the top AI tools as they’ve debuted. In 2023, that meant writing a ton of explainer guides for each one. Here’s the full list, if you’re in need of a one-stop shop for all your AI researching needs.

The big and not-so-big AI models and projects to know about:

• ChatGPT — our beginner’s guide to the OpenAI bot
• Microsoft Copilot — Bing’s generative AI assistant
• Amazon Q — a generative AI-powered assistant
• OpenAI Project Q* — a clandestine AGI Superintelligence project
• Samsung Gauss — Generative AI that produces text, images and code
• Anthropic’s Claude AI — another generative AI from a San Francisco startup
• Einstein Copilot Studio — even Salesforce has an AI assistant for your CRM needs
• Grok — Twitter’s sarcastic AI, available for paying users at $16 per month

Plus, don’t miss our guides to the most recent updates and new functions, from our how-to covering Google’s new AI image generator feature to a ChatGPT update that adds voice command and image input.

Availability: How to Access Gemini

Once you’re fully versed in Google’s new AI tech, you can try it out. You won’t have to take a waiting period or sit through a beta testing phase, either: Gemini Pro is available now through the Bard chatbot website.

It’s only in the English language version of the platform, but it’s rolling out to over 170 countries and territories, so most people will be able to access this version of Gemini. The same is true for OpenAI’s chatbot, which has been available to the public on its dedicated website since its November 2022 debut.

However, both Gemini and ChatGPT reserve much of their powerful features for paying users, whether for ChatGPT Plus or one of Gemini’s other models, Ultra and Nano.