Is ChatGPT-4o Actually Better Than GPT-4? OpenAI's Newest Flagship Model and Its Capabilities

  Рет қаралды 17,759

Corbin Brown

Corbin Brown

Күн бұрын

OpenAI introduces GPT-4o, a faster and more capable successor to GPT-4, offering significant improvements in text, voice, and vision interactions. GPT-4o enhances language processing, real-time conversations, and image understanding, making it a superior choice over GPT-4 for both speed and quality.
SUBSCRIBE for more! 👉 bit.ly/3zlUmiS 👈
Introducing GPT-4o and more tools to ChatGPT free users
openai.com/index/gpt-4o-and-m...
automate everything. 👇
linktr.ee/webcafe
AI Newsletter [FREE] ☕
aitraining.webcafeai.com/joinus
-------------------------------------------------
➤ Follow @webcafeai
• ✈️ 2nd Channel: / @corbinwander
• 🐦 Twitter: / webcafeai
• ⛰️ TikTok: / webcafeai
• 🥾 Instagram: / webcafeai
• 🎧 Bräunlich: / braunlich
-------------------------------------------------
Key Takeaways:
✩ Enhanced Performance: GPT-4o surpasses GPT-4 in speed and quality, providing better interaction with text, voice, and vision.
✩ Improved Language Support: The new model supports over 50 languages, enhancing accessibility for global users.
✩ Broader Availability: GPT-4o is being rolled out to ChatGPT Plus, Team, and Enterprise users, with limited access for Free users and higher message limits for paid users.
▼ Extra Links of Interest:
🤖 AI Courses
aitraining.webcafeai.com/store
My name is Corbin, an AI developer entrepreneur behind the vision of Webcafe AI. Together we will build digital ecosystems. ☕

Пікірлер: 47
@sbowesuk981
@sbowesuk981 15 күн бұрын
An issue with your approach to this test, is you're focusing almost entirely on text to text, the way a user would use GPT 3.5 of 4. The whole point of GPT-4o is that it has been built from the ground up with "omni" capabilities, i.e. it can natively and seamlessly reason across text, audio, and vision. Prior models including 4 Turbo can't actually do that natively (instead using clunky bolt-on solutions), hence why 4o's multi-modal performance is vastly improved. By testing mostly text to text, you're missing the point of this new model why it's actually a leap forward.
@xAnonymousComedia
@xAnonymousComedia 26 күн бұрын
I have a feeling they made the classic chatgpt4 slower and less good to make the new one shine.
@user-oo3fg1ww5w
@user-oo3fg1ww5w 26 күн бұрын
I agree with you.
@angreeee
@angreeee 7 күн бұрын
it actually works faster for me than it used to. They tricked some suckers to use 4o (set it to default) and that new version is not so demanding when it comes to resources. Mor eresources for us - gpt4-ers
@mitchjames9350
@mitchjames9350 3 күн бұрын
I hate 4o OpenAI needs to listen to customers
@BeyondMintReviews
@BeyondMintReviews 27 күн бұрын
THANK YOU FOR USING YOUR HEAD WHEN CONSIDERING A HUSKY! So many people adopt huskies and don’t consider the weather where they live. Thank you thank you thank you!
@thearchibaldtuttle
@thearchibaldtuttle 25 күн бұрын
you could have used chatgpt to tell you how much time it required to give you the answer
@-lunte.
@-lunte. 27 күн бұрын
at 4:50 i could have sworn you were about to say im thinking a rooster! great video dude thank you
@dpdegeus
@dpdegeus 27 күн бұрын
Hey @Corbin, thanks a lot again. Love your channel Tip: I found out that since the update the token limit in Zapier is taken much more strictly. Got the same insights?
@webcafeai
@webcafeai 26 күн бұрын
Thanks for the info! I plan on making more Zapier-oriented videos later this month, so I will check this out.
@brianhopson2072
@brianhopson2072 28 күн бұрын
This is why I like your channel, Corbin.
@webcafeai
@webcafeai 27 күн бұрын
🤝
@luke2937
@luke2937 22 күн бұрын
I honestly feel like gpt-4o isn't that much better in anyways that 4 at a core logic level when just typing to it (I'm writing code and developing a product) Especially when you get into a converstion it often repeats entire outputs rather than the segemented bit I'm talking about, 4 never really did that for me
@sbowesuk981
@sbowesuk981 15 күн бұрын
Depends how you're using each model. If you're using 4o's multi-modal capabilities, it's actually vesting better to 4 Turbo, but text to text, they're both ultimately GPT-4, i.e. both do things you might not want them to do if left to their own devices, so to speak. If 4o is often doing something you don't want, there's nothing stopping you adding a custom instruction to nudge it towards what you want.
@luke2937
@luke2937 13 күн бұрын
@@sbowesuk981 I interact with 4o the same way as 4, specifically as I'm writing code and saas products. Imo 4 just get's it better and provides better responses.
@triste6202
@triste6202 27 күн бұрын
really useful and clear!! liked!!
@OmnipotentCEO
@OmnipotentCEO 7 күн бұрын
Hello Corbin, concerning the coding potion in the 7 minute realm, perhaps AI can do it that way now and we are thinking with our human brain, as far as data organization, how long until AI confirms totally, something is what it is and clears it fully for use on demand? the approach perhaps? is this another branch of AGI you are experiencing without realizing it? all of us I mean?
@user-cv9cd4sq2n
@user-cv9cd4sq2n 28 күн бұрын
So what just like everything else we don’t get to use it but they do
@Designsecrets
@Designsecrets 23 күн бұрын
He said it took ie 10 seconds, and I literally blinked and counted 3.......unsure how this works now
@StevenZhang-vr6yw
@StevenZhang-vr6yw 20 күн бұрын
its so much worse. It doesn't even do certain analysis correctly. The time reduction often means mistakes
@yaroslavzdanovskiy5704
@yaroslavzdanovskiy5704 19 күн бұрын
Agree, it sucks at deep thinking tasks miserably!!!
@angreeee
@angreeee 17 күн бұрын
100% agree. 4o is a regress, not progress. IT has been made to unclog their machines, not to improve response quality.
@AlestroBakai
@AlestroBakai 26 күн бұрын
Does anyone know what the plans are for the ChatGPT 4o mobile app? Yes, it updated and lets you use 4o now, but the app doesn't allow the things they showed in the demo, like giving it "eyes" where it can see traffic, hands, people etc. i can't find any information...
@YannisBang
@YannisBang 25 күн бұрын
That sadly isn't rolled out yet, but Sam Altman said it will be coming in the next few weeks.
@AlestroBakai
@AlestroBakai 24 күн бұрын
​@@YannisBangthank you!
@rokoi3
@rokoi3 17 күн бұрын
The coding test is kinda silly, all that's available when you right click on the page -> "View Page Source". To those learning coding, better to paste that page source into GPT and ask it to explain things step-by-step
@DavidM_
@DavidM_ 10 күн бұрын
He wanted to test coding abilities, not how to use GPT in learning process.
@thecollector4575
@thecollector4575 27 күн бұрын
How do I use this?
@AviynOfTheLord
@AviynOfTheLord 20 күн бұрын
Either the site on computer or the app
@thecollector4575
@thecollector4575 20 күн бұрын
Do I need premium
@falklumo
@falklumo 16 күн бұрын
In my tests, if you ask anything which requires some stepwise thinking, GPT-4 wins over GPT-4o.
@angreeee
@angreeee 17 күн бұрын
gpt-4o is a disaster: 1. not listening to instructions 2. repeating same thing all the time 3. doesn't follow requests to make answers shorter 4. makes the same mistakes over and over again 5. is unable to focus on smaller snippet of code and keep adding unnecessary context even is specifically asked not to 6. produces long unreadable answers each time a bit different, often in conflict with previous responses. less is more and 4o doesn't understand that. They also broke gpt-4 a bit, but it is WAY better than 4o. gpt-4 also produces now long answers, but if you ask him for shorter ones, it usually listens. (i now almost always add "laconic answer please" to avoid 4-page spam). We all came back to gpt-4 after a week of nightmare. (i mostly use it for coding)
@GrounddevLF
@GrounddevLF 10 күн бұрын
thats been my experience so far!
@REASONvsRANDOM
@REASONvsRANDOM 26 күн бұрын
You don’t know whether or not nicely is a word? Hah
@JJ.R-xs8rf
@JJ.R-xs8rf 23 күн бұрын
I literally know nobody among friends and family who use it, most actually hate AI. 😄
@melvinjongen1697
@melvinjongen1697 27 күн бұрын
They still dont get things wright
@_Human_being
@_Human_being 27 күн бұрын
Can't tell from this video. 4:39 in and no fact-checking or quality control check on the info provided. If it still can't get facts straight or provide accurate, holistic summaries, then it's just as limited as 4.
@SierraWater
@SierraWater 28 күн бұрын
Yes, you're all going out of business. FYI. 😂
@user-nj1zu2nf1x
@user-nj1zu2nf1x 26 күн бұрын
why am i paying for a subscription
@REASONvsRANDOM
@REASONvsRANDOM 26 күн бұрын
Loyal subscribers will be given an extra slice of bread each month by the merciful overlords
@senzokai3156
@senzokai3156 26 күн бұрын
In our case, we have a higher use limit, which is well worth the cost. If you're a paying subscriber, using the free plan will just piss you off because you'll have access to the new features but that will be severely limited. This move is aimed at shifting more free subscribers to paid by giving them a glimpse of what the paid plan offers.
@falklumo
@falklumo 16 күн бұрын
What a superficial test. If you make a video, why not spend the time to actually read and evaluate the content?
@melvinjongen1697
@melvinjongen1697 27 күн бұрын
Iv tested it not al to impressed 😂
@Kapreski
@Kapreski 15 күн бұрын
ChatGPT-4 is as bad as a stubborn seven-year-old who insists on repeating the exact same mistake!
@Iguana93
@Iguana93 12 күн бұрын
So GPT 4 summed up something different from the PDF than GPT 4o... "You should make a judgment on which is more relevant..." Geez. It's almost like you maybe should use your brain and read the effin PDF to make your own conclusion and not be a dimwit who wants to have "their work done easier" (meant generally, not as an insult to the creator of this video). I so hate this entire trend and what it stands to do to how people conduct themselves... LLMs are not intelligence, they are unpredictable word calculators made to FAKE human output, not to reason and reliably process information. Anyone asking them to perform your cerebral job for you so that you can sit back is asking for big trouble.
@GrounddevLF
@GrounddevLF 10 күн бұрын
it is god awful spend most my time arguing with the nonsense over the simplest things and when it does work its the most unhuman copied ranting over an over 4 and 3.5 have been wayyyyyyy better
26 Incredible Use Cases for the New GPT-4o
21:58
The AI Advantage
Рет қаралды 673 М.
9 Insane Things You Can Do With the New ChatGPT-4o Desktop App
10:45
New Gadgets! Bycycle 4.0 🚲 #shorts
00:14
BongBee Family
Рет қаралды 15 МЛН
PINK STEERING STEERING CAR
00:31
Levsob
Рет қаралды 15 МЛН
You’re using ChatGPT wrong
9:31
Jeff Su
Рет қаралды 330 М.
GPT-4o is WAY More Powerful than Open AI is Telling us...
28:18
MattVidPro AI
Рет қаралды 246 М.
GPT-4o vs GPT-4: What's the difference?
5:08
Steal These Thoughts!
Рет қаралды 9 М.
GPT-4 Makes Old ChatGPT Look Like a JOKE!
12:56
Nick Chapsas
Рет қаралды 426 М.
ChatGPT 4o vs. Gemini 1.5 Pro - Ultimate Head to Head Comparison!
12:36
ChatGPT 4o vs Gemini 1.5 Pro - A Huge Gap!!
9:56
In Depth Tech Reviews
Рет қаралды 361 М.
Is the AI bubble popping?
19:48
Synapse
Рет қаралды 176 М.
GPT4o and it's new ChatGPT Omni Desktop App for Web Design
8:04
Codex Community
Рет қаралды 21 М.
GPT 4o vs Claude 3 Opus TESTED: Everything You MUST Know!
9:59
Unveiling AI News
Рет қаралды 8 М.
NEW GPT-4o: My Mind is Blown.
6:28
Joshua Chang
Рет қаралды 882 М.