ChatGPT 4o vs Gemini 1.5 Pro - A Huge Gap!!
Вставка
- Опубліковано 15 чер 2024
- SHOCKING RESULTS!
ChatGPT 4o by OpenAI is much more intelligent when compared to Gemini 1.5 Pro, Google should step up the game.
#chatgpt4o #gemini #ai #comparison #openai
*Chapters*
0:00 - Intro
0:21 - Asking for help
3:03 - Solving puzzles
6:;33 - Humor
9:14 - Final conclusion
For Donations:
paypal.me/indepthtechreviews
75 Exclusive Wallpapers on Patreon (as low as $1):
www.patreon.com/posts/15-excl...
Related Videos Playlist:
• Virtual Assistants (Go...
Follow me on:
Facebook: / in-depth-tech-reviews-...
Twitter: / idtechreviews
Instagram: / indepthtechreviews
Amazon Affiliate Links:
Amazon US: amzn.to/3AOk4MB
Amazon UK: amzn.to/36uXJFM
Amazon UAE: amzn.to/3k9LLcv
Amazon Canada: amzn.to/3kg0aUy
Amazon Germany: amzn.to/3hZ4zbT - Наука та технологія
I’m starting to see why Google didn’t do a live demonstration
Google paid Reddit to use their users comments as a training data. Garbage in = Garbage out. There's absolutely zero surprise here.
idk if the maker of the video is paid by openai but i just tried the exact first two questions on my own and Gemini did as good as GPT...
Edit: Tried the IQ Test as well... It did the exact same thing as GPT
@@nero-fv6db I dont believe you i use both alot and gemeni seems like you or talking to a child it way too over protective and lazy. Either it says it just a language model so it cant answer or just give a one word incorrect answer as in the video
@@user-kf8dz7vu2zUA-cam does not allow links or pictures so unfortunately I can't prove it to you but I'm not lying Gemini did as good as gpt and was nowhere near the performance shown in the video...
@@user-kf8dz7vu2z you could just try it yourself...
6:12 The way Gemini answered "21" reminded me of how "beetle juice" answers 🤣🤣🤣
what's nine plus ten
It sent me into space laughing, that 21 was so unexpected.
reminded me of 42 from the hitchhiker's guide to the galaxy
Vro did the 21 savage adlib 💀
Not me who is saying it's fast
gpt: gives you the answer
Gemini: "find it yourself"
Gemini: "I don't know bro, I'm not google!"
This is why Google is so crap. Sundar needs to be fired. They had so much data and so much brains and yet then did nothing interesting with it and on the other hand google search is basically dying now.
@@user-cg7gd5pw5bhuman: but you are.
I'm just a language model, so I can't help you with that.
Lmao 😂
The issue is that gemini is not bad but it's rushed product Google really needs to take some time and then release it
Google has lost its leadership in ai. I don't think there is any other reason. Plus there are focusing on quite insignificant aspects of chatbots like token length.
@@naammekyarakhahai8952 for processing large amounts of data it could be better, no?
I think Google has all the resources and the infrastructure to catch up. They've Google search a vast knowledge pool to draw from
I completely agree with this, although I think it goes a little bit deeper in that they just don't have quite the same talent as openai.
Did you guys not watch Google IO? Lots of valuable AI integrations across search that will change the way we use Google search.
I'm less concerned with asking Gemini to do math and look at what's happening on my screen. I'm paying more attention to how it can be used professionally like document/email creation, sales/spreadsheet analysis, paid ads analytics, marketing campaigns, etc.
ChatGPT has breathing sounds, like when we talk. We need to breathe before the next paragraph, and it's amazing.
which virsion tho? cuz the new voice of GBT 4o OpenAi didn't realese it yet..
@@lostinspace2977 The model behind speech is called Whisper
6:02 The correct answer is actually 39 (because the person is wearing both airpods and a watch). Hence, it's 5+(5+10+2)x2=39. 😊
I was actually looking for this exact comment. Glad someone caught this before me lol
good eye. It's so easy to miss the small details.
@@jimmie_gso true I totally missed that
Or it's 5+17x2 = 44. (as no brackets are in the picture) 😜
Even without brackets you allways multiply first@@MacXpert74
What OpenAI is doing seems almost magic. It’s crazy how good their team is
at this point only thing i am worried is this level of ai "intelligence" is plateauing instead of exponentially increasing.
that this is pretty much the ceiling of smartness its going to be in our lifetime. that its "pretty clever and fast" for logical deductions and guessing from pictures, but thats it, that it wont change the world for much better one bit.
But their product has caused serious problems within the management layer of the company.
@@Redmanticorewe don't need more. I don't need more. Why would i. I still want to work instead of ai doing everything for me.
@@Redmanticore…. Based on what? 😂
@@mitchell10394 based on his copium
Gemini at job interview:
- your CV says you are extremely quick at maths. What’s 5x3?
- 21
- that’s not even close
- yeah but it was fast 😎
😂
this is so funny
2 + 2 is 21, quick maths
😂
But you have to say it robotically and super bluntly 😂😂
6:12 "This is the quickest wrong answer I've ever seen" 🤣🤣🤣🤣
Need this on a Billboard!!!!
Google is in such an embarrassing state right now 😂. Imagine shifting your whole product line to ai, just to be a worse option than the ai everyone already uses
Well the app is lagging behind since its different from browser Variant unlike chatgpt.
The browser variant of both normal and advanced versions gives better responses than what's shown in the video.
In some cases, the advance version outperforms chatgpt.
What do you mean whole product line? It's its own thing. Secondly, it's a new space with lots if growth.
I just think it's a silly remark to make... Either way it's always good to see competition.
Don’t forget, google has the very powerful tpu
@@itsjustme373 And they are asking for paid membership
What about apple. They are cavemen in these, area
ChatGPT 4o is lightyears ahead.. Gemini is still eating "rock per day"...
Agree
Keep daydreaming
Gemini and most google services are worse than competitors cause they lack consistency. The features of one service is not properly communicated to the others. An example of this is Circle to Search. It is a blatant rebranding of Assistant Screen Search yet lacks many existing features. This video shows yet another example, with gemini failing to perform algorithms that google lens and google assistant can already do. Google just keeps making more and more services (then abandoning them months later) when they should instead focus on building upon their existing services
agree
what does circle to search lacks from assistance screen search, you sound butthurt that iphone doesn't have circle to search ever since samsung announced it in February. you haven't healed yet?
@@David-Zitawhat are you on about? Assistant Screen Search is made by Google not Apple. It is the predecessor to Circle to Search which is why I'm comparing them. Assistant Screen Search has Translate, Search, and Work tabs which are missing from Circle to Search. Assistant Screen Search has auto subject detection, making it faster than Circle to Search. Assistant Screen Search also has onscreen QR and bar code scanning capabilities. It is literally superior to Circle to Search in every way asides from fancy animations. Emad literally pointed these out in his reviews. Test both and see for yourself before making these uneducated replies
@@David-Zitahe didn't even mentioned apple and u r assuming he's an iphone user
@@David-Zitahuh wtf are you on my guy who mentioned apple
You are better than Google at informing us about Android and its perks and cons ❤
It's rare to see honest & professional yet simple and understandable comparison & reviews like yours
Thank you brother 😊
I usually just put my prompts in both ChatGPT and Gemini because sometimes one is better than the other
Same
I've been doing this a lot lately, Poe is great for this as it let's you use a ton of LLMs in one app
Since you know, is ChatGPT good at getting Internet results? If it is, I'm literally gonna be switching from Gemini
@@Vex_736 To be perfectly honest, Gemini is way better at that
@@Vex_736 ohh YES it is
It's so accurate of the Gemini experience when it goes out of the way of explaining a joke to tell you it's dangerous to drive in your socks.
06:10 I can't stop laughing 😂😂bro is tired
Why do people replace he she it with "bro" instead of speaking normally? Why not just speak like an adult?
@@nekemli2622 cause it's boring ?
@@jasonstatham483 its immature and cringe and sounds like youre a tik tok user if you say it
@@nekemli2622bro it’s not that deep
@@nekemli2622 You using word "cringe" bro, you're not better. Also, what do you recommend to change word "bro" then ? Like : dude, gal, nigga, fellow human ?
It is sad to see Google lagging behind in the field of artificial intelligence, I hope they learn a lesson
It’s not sad to see; it’s great. Google is already an insufferable monopoly. No one wants them to monopolize AI and ruin it for everyone.
The Gemini app is lagging behind. It's different from the browser variant unlike chatgpt.
Even with normal Gemini on the browser it gives better results than what's shown in the video.
Nothing sad here, fuck Google, they are an evil monopoly company. There is no "don't be evil" about them anymore.
Google isn't worried because they know that in the end, they have the upper hand.
You want one and the same company succeeding at everything? That's fucked up. Look at Intel owning the chip market for decades and look what they provided for the consumers.
I just went from ChatGpt 4o to gemini 1.5 pro. Regretted it instantly. Gemini is very dumb compared to Gpt 4o
Another amazing video! Looking forward to a series based on this, and with the upcoming "Siri improvement" rumors we keep hearing about. Well done!
Him: What's the answer to this test😊
Gemini: 21🗿 (no context)
You are the only UA-camr I click as fast as the notification enters.
same
Same
same here =)
Same
Gemini’s response to the car prompt sounds like my dad!! “It is over there, why don’t you go and look for yourself instead of bugging me”
Seems that Gemini was just not in the mood, „… you need to watch the video!“ 😂
Especially the way it said "better" 😂😂😂
Google is not using its own resources and strength for Gemini which makes it utterly unusable and behind. Even the Google assistant which is older is still ahead of Gemini
Google Assistant ahead of Gemini? Not in my world, you can barely compare these two imo 😀
@@Daniel-jv9ft for my every day use, Google assistant surpass Gemini significantly. I may use it for rephrasing and some debugging, but for managing my smart home and checking my schedule... Assistant is still way ahead
@@Daniel-jv9ft Gemini feels like 2007 tech it is so bad.
The blunt "21" floored me 🤣🤣🤣
Man It's literally 21 and no one gets it 😂 lemme explain:
In the 3rd picture 2 watches=2 so that means 1 watch=1. And in the last one the person is wearing 2 airpods (10)+1 watch not 2 (so 1). So because we also add 1 airpods instead of 2 (we add 5)
It's 5+[(5+10+1)×1]
Priority to the parenthesis then the multiplication because of BODMAS. So it's 5+(16×1)
=5+16
=21 !
I'm honestly surprised Gemini got it right and that no one in the comments gets it
@@MainInternetUser In the first pic a single airpod x 6 (since there are 6 individual ones) is 30. So one airpod is 5. In the second row, we know that the two AirPods add up to 10. So whatever the humans are, they have to add up to the remaining 10. So 10/2 is 5. One human is 5. In the 3rd row, we already know a human is 5. So the 4 watches need to add up to the remaining 8. 8/4 is 2. So 1 watch is 2 (not 1). So the final equation should be 5 + 17 2. The 17 is the human + 2 AirPods + 1 watch (5+10+2). Order of operations says multiply 17x2 to get 34 and then add the 5 which is 39.
@@KelabeteWe can express this as a system of equations. Let x be a single AirPod, let y be a person, and let z be a single SmartWatch.
So, the system is:
3 * 2x = 30
2y + 2x = 20
2 * 2z + y = 13.
Find x + y * z.
Solve for x:
6x = 30
x = 30 / 6 = 5.
For y:
2y + 2(5) = 2y + 10 = 20
2y = 20 - 10
2y = 10
y = 10 / 2 = 5
For z:
4z + 5 = 13
4z = 13 - 5
4z = 8
z = 8 / 4 = 2
So, summing all of the values, we get: 5 + 5 * 2 = 15.
@@loxitto this is sound, but the last equation is isn't simply airpod + human x watch. It's airpod + (human wearing two AirPods and one watch) x watch.
So with your equations for finding x, y and z, we know both x and y = 5, and z=2. Plug that into the last equation, it would look like x + (2x+y+z) x z or 5 + (10 + 5 + 2) x 2=n
Order of operations says do the brackets first (17). Then multiply the sum of the brackets by 2 (34) and then add the 5 (39). I think ChatGPT struggled to see the minor details on the human (which is meant to trick a lot of people)
@@Kelabete Oh, you are right. Even myself didn't see that detail.
I'm very happy that you made a new video with this topic again with premium models. Also, I couldn't find the old video on your channel where you compared the GPT 4o to the free regular Gemini AI. Kudos to you mate 👏
Great review with great examples. Please continue to do reviews of different models like this.
What a great video👍🏻 Was a pleasure to watch
Google trying to compete in the LLM race is basically like Orkut, Google+ all over again 🤣
This is very interesting, I took screenshots of the examples and uploaded them to Gemini, responses were completely different, however there is no justification for answers in the video to be incorrect. Also from my personal observation, after rollout of Gemini 1.5 the responses did become shorter and less informative which is quite sad
Gemini = Quick Maths! Ain't correct, but DAMN is it quick! 😂
Thanks for all your great videos.
This video really shows how far behind Gemini really is compared to the competition, I can't believe they charge a monthly fee for this.
Great work. Keep it up please.
also please do comparison test between Gpt 4o vs co pilot... any differences?
copilot is using normal GPT4 right now which is less capable than gpt4o but copilot is soon getting an upcoming update with the new upcoming AI version of windows and will also use Gpt 4o as base. still even Normal Gpt 4 is better than gemini 1.5 pro
Beautiful comparison! Best channel for all things Pixel.
Awesome video as expected. Thank you sir.
Gemini is like the home tutor whom we pay to solve problems but instead he tells "Do it yourself"
Yeah Gemini is somtimes just straight up stupid (sadly) and I'm only using it, because I got 2 month Gemini Advanced for free 😂
How did u get it for free
@@nuskyahmad6375 I got it because I was previously subscribed to Google One, but I already know I'll cancel after the 2 month and return to my previews plan.
first time I've seen this format. Subscribed.
Awesome video thanks! I'm trying to compare which to subscribe to and this helps! Any plans to compare some others?
6:00 actually it's not 15, the correct answer is 39
How could it be 39
Oh wait the person is wearing the earphones and the watch sorry I didn't notice
Yeah I counted it now with that in mind and yea the answer is 39
Its 156
sorry, but why is it 39? I count 24: One EarPod (5) + Person (5) with two earpods (10) and with one smart watch (2) Plus one smart watch (2). 5 + (5+10+2) + 2 = 5 + 17 + 2 = 24. Am I wrong?
شكرا لك على المقارنة المنصفه كـ عادتك بدون تحيز وايضا فيديو جديد بالفكرة ❤😊
Thanks for this review video very good comparison. I was pondering to sign up for ChatGPT 4o or Gemini Advanced until your this video helped me make the right decision that ChatGPT 4o is the one to go with! 🎉👍🏻
Thank you for explaining it so nicely, it will help me a lot
Gemini seems like a student rewriting a question as the answer!
Fr fr. Lol
Even though ur a Google fan, u keep it very honest n real. This is very bad for Google. Especially if Apple gets ChatGPT to run Siri or whatever. I use both platforms this can reverse the idea that pixel is the smartest phone n iPhone is pretty dumb.
I will subscribe to this channel just because it said the answer in the beginning. No bullshit! I liked it
Excellent job! Even I did not answer to all of question correct, but GPT was ahead
Can you test GPT vs Gemini vs copilot ?
Copilot is built on GPT-4, though it's not GPT-4o
@@TheStopwatchGod Actually it was announced at Build that CoPilot is being upgraded to GPT-4o.
Would love to see that test. Copilot quality collapsed after they “upgraded” to GPT 4 Turbo, I find it useless now while for over a year it was the best, because it could also search the web and incorporate those results.
@@jamesdax yea but with the upcoming Ai windows update
Can you test ChatGPT-4o vs Claude AI?
Claude is cringy and dumb af
why dont you test it if you want to know and make a video? they cost what, 20 bucks each? ask it similar questions
" Claude 3 Opus, our most intelligent model €18/month + VAT "
great comparison, thank you! :)
Very cool comparison! Thank you very much. ChatGPT 4o is really impressive :)
I thought google was trolling for a second with that "21" answer. Like that decade old 'What's 9 + 10?' meme.
Considering that it's trained on human data... It's not unreasonable to believe that _some_ of that answer can be attributed to the existence of that meme.
Google paid Reddit $60 million to train their AI on Reddit user comments. That cesspool is full of trolls and fake info. It's no surprise that the Google AI is shit.
Yeah Gemini is way less smart than GPT 4 and Copilot. I honestly never use Gemini over Copilot because Gemini is just way dumber. It doesn't code/program good or anything. I find GPT 4 and Copilot miles ahead. Google needs to step it up.
I think it's important to note that this is only testing scenarios where the input is an image. The general intelligence of the LLMs might be different in text-only conversations.
It would be also interesting how both tools perform in data analysis tasks dealing with plots, data frames and with scientific publications
"Only a few will understand what's going on here". What if we all don't understand and it's Gemini that gets this highly sophisticated joke?
Gemini is Bing 😂😂
Bing's copilot is the same chat gpt
Once I asked Gemini about Pixel phone and got wrong technical information. At that point I understood how Gemini is unfinished product. Even in presentation Google people were talking about unknown terminology without clear demonstration. ChatGPT's was perfect.
Best comparison ever.. you rock!
I will cancel Gemini Advanced.
This is why I cancelled Gemini
Really appreciate your efforts😊
New subscriber
Gemini can watch the video now.
Summary: ChatGPT is too chatty, and Gemini too blunt.
Job well done. Thanks for sharing
Can you make a coding/development related comparison, like sharing a file/block of code and asking them to explain it line by line, or sharing, then make the necessary modificatios to implement a specific feature or something.....? That'd be extremely helpful
ChatGPT excels in natural language processing, offering unparalleled conversational intelligence.
Google Gemini stands out in solving complex mathematical problems with precision and efficiency.
எண்ணென்ப ஏனை எழுத்தென்ப இவ்விரண்டும்
கண்ணென்ப வாழும் உயிர்க்கு.
But it got the math question wrong?
Gemini is even worse at maths actually. Lol
Great choice of subject to make a video!
I too experience it the same way. Many times I try to use gemini, but her replies make me come back to chatgpt again. Both models are in their early forms, but chatgpt indeed feels more user friendly.
This was an amazing revelation between the 2 popular AI tools...could you plz make a similar one and include Claude and perplexity AI too...
Very nice review, love it 👍
Great comparison ❤
Had to click as soon as I saw this upload
Yeah, Gemini is really funky. I can't even get it to consistently set alarms or timers as sometimes it claims I need to upgrade my plan, sometimes it says I need a separate app and sometimes it just works. It also requires additional information quite often when no more info should be necessary.
It seems really dependent on phrasing and if you don't get it right it just doesn't work.
Very interesting. Who would have thought that Google with such a massive data pool to begin with would get it so wrong. I've tried regular free versions of the AI's. Must say now I get why every time in an instruction/request/inquiry with Gemini, I have tease out the info I am looking for with multiple addition inputs post the initial request even when I would have entered the info in first place. ChatGP works perfectly with large initial instructions/inquiry.
And this is the perfect sample why competition is so important
would you do a video about GPT4o vs Copilot?! ot would be insightful ♥️
GPT 4o has helped me a ton. So far, just formatting data has been super helpful and it understands what I want much better than before.
It's amazing that it's going free.
Does Gemini have memory? I have spent time putting my preferences and interacting with ChatGPT, and it knows I prefer fuller responses. Wondering which you use more and if that could be a factor in their interaction
Very good, detailed video.
Great video, keep it up!
Amazing video it really shows how drastically far ahead openAI is of its conpetitors
i did my own comparison by integrating them into my product development workflow. This includes product concept design, marketing, frontend dev (flutter), backend api dev (python and fastapi), postgres data modelling, and more.
Gemini very quickly fell to pieces once i started using it deeply. It wouldn’t fully understand the request, it wouldn’t remember ways i previously asked it to answer me going forward, and downright got answers very wrong even when it needed to search the net for answers (as it would refer to very outdated online articles).
My backend runs in GCP, but gemini - at least right now - is so far behind and wrong so often, i just have to go with OpenAI for my systems AI features.
If we freeze ChatGPTs dev in time, i’d put Google at least a year behind them. But that’s the thing, OpenAI is improving at an incredible rate.
Thanks for this excellent video.
just tried same questions on gemini advanced web; it seem to give same or on par answers with chatgpt 4o through??
I don't know why the app is way dumber. It should not matter, the only difference between the app and the web should be the UX
Does this already have the features announced at Google I/O? Because there they showed Gemini basically watching the video for you
i used it in unreal engine 5 i think blueprint image questions can really mix the difference and make it harder on the chats but gpt4o came out near flawless i even created a complex blueprint formula that gpt4 failed with in the past and it passed with ease.
1:30 "Comparing the Tyotta Camry" أحلى مسا عليك يا صاحبي
Some of the answers you get with this model, really depends on the query, I had asked both some of the same questions before I even watched the video and got different answers from the ones in the video, and in some cases, my answers are the same from both AIs.
very informative vid.. thanks
Can you compare gpt 4o to the most advsnce lamma 3 ai model with 400b parameters
I m pixel 8 pro india user
Can I use chat gpt or gemini in my pixel without any third party app download natively or do i hv to download apk separately to get these features work
Oddly enough, and it may just be my experience so far, but i have been trying both for odds and ends in my business and 70% of the time Gemini has been giving me better answers. I use mine mostly for research, which might be why.
"It looks like gemini hallucinated on this one" that got me 😂
May i know why the AI meme understanding comparison video is no longer available?
I replaced it with this one that includes memes and other stuff too
You asked very good questions.
Oh boy! I felt your pain of Gemini responding with just '21' 😂
Convinced me to keep subscribing to ChatGPT over Gemini. Thanks! 😊