Deciding whether you should use ChatGPT or Google Gemini can be a fleeting exercise. Companies like OpenAI and Google are updating their AI chatbots frequently, meaning that the quality of answers you get can differ from month to month. Additionally, each question you ask an AI chatbot outputs a unique answer, making it more opaque when comparing answers side by side.
In the battle of free AI chatbots, Gemini's larger feature set makes it the one to pick over ChatGPT. Gemini includes image generation and a connectivity with Google's other products, like Maps, while ChatGPT lacks such amenities.
Historically, Google Gemini performed worse than ChatGPT, Microsoft CoPilot, Anthropic's Claude and Perplexity, as noted in our Gemini and Gemini Advanced reviews from earlier this year. But as of September 2024, Google has fixed some of the hallucination bugs that have been plaguing its AI chatbot. In terms of quality of output, hallucination rate and features, the free version of Google Gemini is competing on par with or, in some cases, better than the freely available ChatGPT-4o mini, OpenAI's lightweight AI model that's zippy and has a connection to the open internet. An LLM, or large language model, is an AI trained on massive amounts of text data that can then generate original text like an advanced version of autocomplete. At the very least, both Gemini and ChatGPT Free are connected to the open internet and can mostly answer topical questions.
Let’s compare ChatGPT and Google Gemini so you can determine which AI chatbot is right for you.
ChatGPT Free vs. Google Gemini: What do you get with each?
ChatGPT Free -- which runs on the GPT-4o mini model -- and Google Gemini -- running on the 1.5 Flash model -- are both freely available AI chatbots. Gemini requires you to log in with your Google account, whereas ChatGPT can be used right away. However, ChatGPT says logging in will lead to more accurate results. For my testing purposes, I stayed logged out.
Both chatbots can assist with writing, brainstorming, general inquiries, travel and shopping recommendations. But there are some differences between the two. ChatGPT doesn't feel as restrictive as Gemini, willingly answering questions about politics and current events. Given Google's past snafus, like portraying people of color as Nazis, it makes sense that Gemini has been tuned to be on the side of caution. Gemini can generate images with the Imagen 2 model and allows you to upload images for analysis, a feature absent from ChatGPT Free but present in the paid ChatGPT Plus plan.
Both AI chatbots do have limits on the number of questions you can ask in a single session if traffic is high. In my testing, I didn't run into these restrictions. Unless you're someone who needs to do a lot of data analysis or have an AI chatbot read over hundreds of pages of text, the free versions won't be restrictive for most general questions and inquiries.
ChatGPT vs. Google Gemini: What do you get with the paid plans?
ChatGPT Plus and Google Gemini Advanced are the paid versions of these AI chatbots. They tend to be a bit slower in their responses but churn out more accurate and verbose responses. The paid tiers also allow for higher-end image generation -- and in the case of ChatGPT, you can create images only using the paid version.
Both the paid tiers for ChatGPT and Google Gemini provide access to specifically tuned models, what OpenAI calls GPTs and Google calls Gems. These models also allow you to input a lot more data if needed, with Gemini Advanced promising to one day allow up to 2 million tokens, which is approximately 1,500 pages of text. ChatGPT Plus has a 64,000-token limit, by comparison. There's also the ability to upload audio files and video for analysis and chat with advanced voice modes that sound more human. For basic inquiries, most will find the free versions of ChatGPT and Gemini to suffice. Ultimately, it's worth paying for ChatGPT Plus or Gemini Advanced when you want more verbose responses with higher-level thinking.
How do responses compare?
Google Gemini and ChatGPT Free trade punches when it comes to giving accurate product and shopping recommendations. Both chatbots had instances of giving incorrect answers. In some instances, ChatGPT would give more detailed or nuanced information, whereas Gemini would sometimes be more accurate. Overall, however, both gave good enough information to be adequate for basic inquiries, from shopping to travel.
Gemini hallucinated slightly less, but didn’t always offer the right details
When asking for shopping recommendations on a mid-range turntable available in red, ChatGPT recommended units that didn't even come in the cardinal colorway, whereas Gemini immediately pointed to the few red options on the market. Even then, Gemini recommended one unavailable in red. As for cartridge recommendations, when I specifically asked for ones that prioritized a neutral sound that aims to be an accurate reproduction, both chatbots pointed toward products that produced "warm" tones. The use of "warm" or "smooth" is a common refrain found in cartridge reviews online. It makes sense that both chatbots brought up this language as it's common verbiage used when describing tonal texture, but it really shouldn't have been included in the outputs because the focus should have been a neutral sound, not warm or smooth. And it shows some of the limitations of deferring to AI chatbots for information, which is an average of text found online. Still, both chatbots also highlighted other cartridges that were tuned toward accuracy.
In comparing two of the best OLED televisions for gaming -- the LG OLED G4 and the Samsung S95D -- both Gemini and ChatGPT Free provided an accurate comparison between the two, highlighting their response times and color accuracy. ChatGPT did a better job of capturing specific nuances, like the Samsung S95D's increased reflection blocking. And ChatGPT actually told me the input response times in gaming, whereas Gemini said it was difficult to provide that information.
ChatGPT sounds more authoritative -- even if it isn’t
While each AI chatbot has advantages in handling shopping recommendations, ChatGPT pulls ahead when responding to questions about philosophy. To put it bluntly, ChatGPT sounds like it knows what it's talking about. Gemini comes across as too agreeable, as if it doesn't know the exact answer. But by echoing certain words from the question asked, it can give a general enough answer to pass as adequate.
Gemini is completely useless when asking it questions about the upcoming presidential election, oddly. Asking it to give thoughts on the debate between Donald Trump and Kamala Harris, it said, "I can't help with responses on elections and political figures right now. While I would never deliberately share something that's inaccurate, I can make mistakes."
At the end, Gemini simply said to Google it instead.
ChatGPT, on the other hand, was able to quickly spit out a summary of what pundits had been saying post-debate.
That's not to say Gemini shies away from all topically hot discussions. It openly chatted about abortion rights and laid out arguments on each side of the debate.
Both ChatGPT and Gemini had some similarities in their generative text performance
Weirdly, when I asked both chatbots to write a short story in the same vein as All The Dirty Parts by Daniel Handler, both gave short stories about chocolate and other confectionaries. Handler's book isn't about sweets; rather, it's a blunt and gross account of teenage sexual affairs and perversions with little censorship. This could be an error in the training data, or maybe each chatbot was avoiding getting into the explicit sexual imagery found in Handler's novel.
Maybe that test was too difficult for Gemini and ChatGPT to parse. Instead, I asked them to write a short story in the style of Dr. Seuss. Gemini may have committed plagiarism with its story titled, The Whos of Whoville and the Whos of Where. ChatGPT, on the other hand, wrote a whimsical story about Fizzle McSound. A Google Search of Fizzle McSound didn't produce any relevant results, suggesting that McSound is an original character. At the very least, both stories sounded like they were from the school of Seuss.
Gemini has images even on its free plan
One major advantage Gemini has over ChatGPT is its ability to generate images on the paid tier -- ChatGPT’s Dall-E 3 image generator is limited to the premium ChatGPT Plus plan. Gemini has some limits, like not being able to generate images of people. And it's using the less advanced Imagen 2 model, although Imagen 3 support is coming soon. But at the very least, it's an option available.
When asking it to generate images of kittens eating spaghetti, the results were adequate. There's certainly some AI weirdness going on, like the fur looking too diffuse and with eyes that look like Cheshire Cat from Alice in Wonderland. It's passable but not perfect.
ChatGPT vs. Google Gemini: Final thoughts
After reviewing multiple AI chatbots this year, it's surprising to see how much better Google Gemini has gotten. Compared with ChatGPT-4o mini, Gemini hallucinated less while giving competitive answers. However, Google has tuned Gemini to be a bit too clean, with it outright refusing to answer questions that might get it into hot water from politicians, like not wanting to summarize the most recent presidential debate. ChatGPT didn't have such concerns.
Gemini is also linked to Google's other products, like Maps and Search, meaning it can give you directions if you want to plan a pizza tour around Brooklyn. Granted, it's not without fault. Gemini's Brooklyn pizza tour had me starting in Manhattan and zig-zagging between far-flung neighborhoods that would have led to too much time commuting in the subway. While ChatGPT couldn't instantly generate a map of a Brooklyn pizza tour route, it also had me jumping around disconnected neighborhoods. Ultimately, while the recommendations were fine, it's still best for you to plot out the points on Google Maps yourself.
At the very least, it's handy that Gemini allows for image generation -- a feature not present in ChatGPT Free -- even if it's limited to animals, cartoons and inanimate objects. Google got into a bit of trouble when interpreting historical figures earlier this year.
Still, given that both ChatGPT Free and Google Gemini have comparable performance, I'd ultimately end up recommending Gemini as it has more features. While it can't answer all questions, it does a good job of answering most. And if it refuses to answer something, ChatGPT is only a click away.
Plus, it can generate images of feline culinary hijinks.