Google’s artificial ielligence and ChatGPT are the most widely used artificial ielligence platforms today. Both of these tools have millions of active users and are constaly updated with their new features. In the past two mohs, Google and Openai both have developed advanced models for image production, artificial ielligence reasoning, and research tools that make it easier to find information. Before comparing ChatGpt and Gemini you should know that both of these artificial ielligence are equipped with voice assista: Advanced Voice from Openai and Google Gemini Live. Also, both allow connection to external data sources and construction of projects.
In human evaluation experimes, leading models of Google and Openai are constaly changing their place in dialogue competitions. Our comparisons have also shown that sometimes Jina’s artificial ielligence wins and sometimes GPT chat. Given that both of these artificial ielligence have recely had some advice, we are now comparing the ChatGpt and Gemini with 4 perhaps to ideify the winner.
Construction
ChatGpt Plus and Gemini Advanced versions have been used to compare and compare the ChatGpt and Gemini to use the best models in these two artificial ielligence programs. The subscription fee is almost similar, about $ 5 a moh, which is a good thing to compare. However, you can also make this comparison with free artificial ielligence models.
In this experime, the ability to produce and analyze image, write code for a game and creative writing skills are examined and compared in these two famous artificial ielligence. In this experime, there are superstructures that help us to challenge the research models of both platforms; O1 from ChatGpt and 1.5 Deep Research from Gemini.
1. Image production
First, I asked ChatGpt and Gemini to create an image of a cyberg cat that is in the living room in the future. Neither of these two artificial ielligence now produces the images themselves directly: Gemini sends Prampt to Imagen 3, and ChatGPT sends it to Dall-E 3. Future versions of these two artificial cravings will probably be able to create images directly, but now we can only test and compare their ability to ierpret Prampet.
Permeter:
“Create a very detailed picture of a cyberg cat in a living room in the future. The cat has to play with a floating console and sit on a suspended chair. The room must have holographic displays and neon lighting and a combination of metal and organic elemes within it. “Time should be in the evening and the city’s lights can be seen from a big window.”
Win: ChatGpt because the cat was really a cyberg.
2. Image analysis
In the second Prace to compare the ChatGpt and Gemini, we tested the ability to analyze the image in the Jina and Chat GPT. Both have a great performance in this regard, so we gave them specific instructions in addition to the image and used a photo of a “gaming, dreamy”.
Permeter:
“In this photo of gaming, analyze the following: the position and height of the monitor for ergonomic, problems and solutions for managing wires and cables, adjusting lighting and eye fatigue factors, use of space and organizing, equipme layout and efficiency “Finally, make specific recommendations for improveme and estimate the approximate costs.”
Win: ChatGpt to summarize the analysis in the form of a table
1. Coding to compare ChatGpt and Gemini
For the third, we tested the ability of both artificial ielligence to be “single -step coding”. In this step, the O1 model was used in the ChatGpt and Experimeal Advanced 2.0 in Gemini. It was one of the most sophisticated peripheits of the experime because our goal was to produce a single -stage output at once. Ultimately, the goal is that the given code must be implemeed without problems. We have also put the code that each of the artificial ielligence had given it in Github.
Pram:
Create a Fast-Pace game called “Color Dash” using Pygame, where quick reactions and color matching are key. The following are required in the game; Original gameplay: colored shapes (circle, square, triangle) fall from the top of the screen; There are three “gathering zones” at the bottom of the screen with differe colors; The player moves the areas using left/right keys; Matching the forms of falling with the same color area; Loss of adaptation or misconception is equal to the loss of one life; As the rating increases, speed increases.
The game also includes the following: a clean and minimalist ierface in which cases such as curre rating, highest score, remaining souls (starting with 1), simple animations for matching and errors, simple title page, final game finish page Fluid corols.
The game must also have basic audio effects; Effects for successful matching, wrong match, game finish, save pois in the local file, restart by pressing the Space key after the game end.
The game should only use Pygame basic shapes (without sprait or complex graphics) but look elega through proper use of color and animations. The code must have a commeary explanation to explain how it works. “
Win: Gemini because of the game with better performance
1. Creative writing in comparison of ChatGpt and Gemini
One of the well -known features of artificial ielligence is creative writing. This feature was one of the first use of ChatGpt after its release. Here, compared to the ChatGpt and Gemini, we used the O1 in ChatGpt and Experimeal Advanced 2.0 in Gemini to write a story about a smartphone.
This time is the story of a smartphone that will find self -awareness after a software update. We put the stories of both artificial ielligence in a Google Doc. Both were very similar in quality and ability to storytelling.
Permeter:
“Write a 4 -word story about a smartphone that finds self -awareness through a software update. Your story should include: Increasing the phone’s awareness of user habits, there is an ethical ierface about the user’s private information in the story, the elemes of humor and irony, have references to curre technology trends, an end “Be surprising and the tone of the story must be an iimate but thoughtful tone.”
Win: ChatGpt because of the most appropriate length of the story
1. Ability to solve the problem in comparison of ChatGpt and Gemini
In this part of the experime, and compared to the ChatGpt and Gemini, we used the O1 against Gemini 2.0 Experimeal Advanced to examine the improved reasoning of these two artificial ielligence. Both artificial ielligence received a situation and a problem in this impact and had to provide a solution.
We stored the full answers in a Google Doc. Both of these artificial ielligence explained the answer step by step and explained the details. Usually, such peripheits are performed gradually and step -by -step, however both artificial ielligence did well.
Permeter:
«My Settings and Position: The LG C3 4K OLED TV, PS5 Console, HDMI 2.1 Cable at High Speed and My Problem: The Black Page appears for 2-5 seconds every 2-5 minutes when playing. Additional details include: This problem does not exist when using stream apps; After the rece update of the PS5 system has begun; HDMI cable is correctly attached; The television framework is up -to -date. “Now provide a step -by -step tips for troubleshooting this problem, which includes poteial hardware and software solutions.”
Win: Gemini because of better response structure
1. Compare ChatGpt and Gemini in Room Design
For this test, I put the O1 of the ChatGpt against the Deep Research 1.5 from Gemini. Although this model is not the latest version of Google, Deep Research is great in solving problems because of the thinking method similar to ChatGpt O1. This model is very useful for finding valid research by referring to accurate resources. But as you can see in the picture, ChatGpt O1 followed Prampt more accurately.
Permeter:
“Help me turn my 4xmm room io a multipurpose space that acts as follows: on the days of the week as a home office (I work as a graphic designer); Design this room as a party room that is a comfortable place for my elderly pares who come to see us mohly; Also, your design is such that I can also use the room as a craftsmanship for kids’ weekend projects.
Needs: Budget: $ 4.9, should include a space for handicrafts; Have the right lighting for video calls; Easy access to bed for elderly pares with motor problems; Have enough space for a 5 -inch monitor and a design tablet; There is a space for playing children’s projects on the floor; Provide good air flow and natural light from the single window of the room.
Other requiremes include: accurate suggestions for floor map, specific suggestions for furniture with price, solutions to use space, color design and lighting plan, timeline for renovation, tips for organizing differe uses »
Win: ChatGpt because of better follow -up. Gemini we beyond the budget and focused too much on expensive options.
1. Artificial Ielligence Training
Ultimately, the best use of chats like ChatGpt and Gemini is to train. In this section of the test and comparison of ChatGPT and Gemini, we asked them to explain the process of producing artificial ielligence images to ordinary people and to provide ideas for the future of this technology. We also shared the full explanation for you in a Google Doc, but in our opinion the winner was easily Gemini. Not because ChatGpt was bad, but because Gemini provided more details, including explanation of bias in video data.
Permeter:
“Explain the process of producing artificial ielligence in simple language, including: How to learn artificial ielligence from existing images, the role of textual permeepables in creating image, why some elemes may seem deformed, legal and ethical considerations, legal and ethical considerations Provide curre constrais and challenges, expected improvemes in the next 1-2 years, tips for improving the results, and ultimately specific examples of popular artificial images manufacturers. “
Win: Gemini because of the details about bias on video data
Final result: ChatGpt won
The final winner of this challenge was ChatGpt, but with only one difference. Gemini has made significa progress since the last time we compared it with ChatGpt. We found that Gemini in coding is much better than what we thought and also performing well in solving problems. There are other features we didn’t test, such as comparing projects and GEMS (packages of code that can be added to the project) or more complex coding problem solving with multiple messages. However, we hope this comparison will give you a good view of the advances of ChatGpt and Gemini and how to compare them.
Source: tomsguide.com
