GPT-5 is INSANE! (FREE)

It's official. GPT5 has been released. I'm going to be showing you how to access it today. If you don't currently have it inside chat GPT, no problem. I'll show you some other ways to access it. And this is the biggest release of the year when it comes to AI. So, you can see on OpenAI's official website here, they've said, "Introducing GPT5, our smartest, fastest, most useful model yet with builtin thinking that puts expert level intelligence in everyone's hands. " One of the interesting things about this is Sam Orman the founder of OpenAI Jack GPT says that this is basically like a SAS ondemand AI model and also you don't need to switch between different models. It will just figure out which one is the best. So you can see here they've said we're introducing GT5 our best AI system yet. GPT5 is a significant leap in intelligence. This is better than all of our previous models featuring state-of-the-art performance across coding, maths, writing, health, visual perception, and more. All right. And it is a unified

model. Now, what that means is essentially if you're using Jack GPT, you can switch between these different models like thinking or you've got quick coding analysis, reasoning models, etc. advanced reasoning, it's all a little bit messy. Whereas inside GPT5, you'll be able to access this in one single unified model and it will just pick out which one is the best one for you. All right. Now if we actually have a look at this as well, we have access to GPT5 inside open router. So you can see there are many different models. You've got for example GPT5 nano, GPT5 mini, open GP5 and then GPT5 chat as well. You can also start coding with it inside playground AI. So you can go to the chat here, start using GPT5 directly here and you can code with the API key as well directly. Now, if you want all the notes from today, feel free to get it inside the AI profit boardroom link in the comments in description. Just go to the SP section and all the notes from today will be inside there along with loads of other stuff including daily live coaching and all my best courses. GT5

not only outperforms other previous models and benchmarks and answers questions more quickly, but more importantly, it's more useful for real world queries. All right, so it significantly reduces hallucinations, which is a great move in the right direction. Interestingly, they talk a lot about health here. When I'm looking through the page writing code inspir health is a new sort of description of the model that I've not really seen before. So, they obviously made some advancements. It's going to be interesting to see how that's used for health and that sort of thing in the f. The other interesting thing here is it's good at front end generation, right? So, for example, if we have a look what people are generating with it, people are like oneshotting games. It's pretty crazy. If we have a look at this tweet from Alen, he said, "What's truly impressive about GPT5 one shot in this game is in the graphics is the flawless prompt adherence and constraint handling. A fully functional interactable game generated in a single pass. This level of instruction following and code generation is wild, right? And you can see an example of what that looks like right here. So you can basically like oneshot games on the front end, that sort of thing. You can see how it performs in terms of numbers. So, if we look at the different versions of GPT on the scale, it's off the charts, isn't it? It's just it's ridiculous compared to other people. Let's keep going through now. See what we got. Sam Wolman is tweeting about it

as well. So, he said, you know, when you get access to GPT5, try a message like, use Beatbot to make a sick beat to celebrate GPT5. And then you got like this SAS tool that's just been created. You can see how quick it is. It created that in literally 6 seconds which is absolutely wild. And the UI is pretty nice as well. So it can like create games and SAS and that sort of thing on the front end too which is pretty exciting. And this literally just came out. I think they literally just did a live stream like 2 hours ago as I'm making this video. Right. You can see here. Yeah, the announcement just came out 1 hour ago. This is brand new. This is fresh off the press and most of the world has not even seen this yet. So, if you're watching this, you're probably in the top 1% of the world seeing the latest releases from AI. The other thing I was going to say is this has different models. So, we got GT5 Mini, etc. And GT5 is the strongest coding model to date. So, you can see the differences in the charts right here. And this is the other thing as well. So, they've got different personalities inside the chat. So, you can, for example, use synic, robot, listener, or nerd. And then you can customize those in your chat GPT settings. You can also switch between like research text only opt-in change anytime blah blah and you've got some examples right there. Right. So that's what it's going to look like once you've got access. Now this is rolling out to everyone around the world. So again I don't have access inside chat GPT directly but I do have access inside open router. All right. So if you want to get access directly just go to open router if you don't have it already released inside chat GBT. Here's another example. So, this was just one single prompt, basically making it the strongest coding model ever when it comes to building stuff out. So, we've got the jumping ball runner. I'm just going to mute that so it doesn't get too noisy. But yeah, it's built this game. This is actually playable as well. So, it's been embedded directly on chat GBT's website and then you can build out and oneshot games and that sort of thing, which is pretty cool. Pretty nice looking UI right there. That's the prompt that we use. We can actually test that right now. So, if we take this problem like and then we go over to open router. Let's just try what we got here. What we got? So, we got nano mini. Let's just try this one instead. So, I'm going to go directly inside the chat here. You can see the price and that sort of thing. We can look at the specs as well in a sec. There's a lot to talk about here. about. Who we got watching? By the way, if you're watching this live, shout out to you. Thank you for watching. Let us know which countries are you watching from. Put a flag in the chat. I'll give you a shout out. So, this looks like it's using the reasoning model. If we just go straight into chat GBT, we didn't ask it to, but I think it will automatically choose between the different reasoning models and then we can come back to that in a second. Right. So, I've used exactly the same prompt from the website for the game and we'll see if it actually does oneshot it perfectly like we've seen today. In a minute as well, later on in this video, we can compare it against different models. See how it goes as well. We got someone from Algeria. I know it. Norway, shout out to you. So creative expression and writing T5 is our most capable writing collaborator yet able to help you steer and translate rough ideas into compelling resonant writing with literary death blah blah. Let's see

what else we got here. So this is I'm really interested about this. Like why are they talking about health so much? I know they did have a specific model for health. I didn't think that would be integrated with GPT5 so much but this is interesting. Compared to previous models, it acts more like an active thought partner proactively flagging potential concerns and asking questions to give more answers blah blah. I wonder if for example some people going to the GP and that sort of thing whether they're just going to turn to chat GPT for the health related stuff. This is interesting right? So for example it's comparing GT40 versus GT5 to write short poem and then you can compare the outputs. Right? So you got GPT4 over here and you got the outputs of GT5's poem over here. This is a lot more poetic than this one, right? It's still using those crazy M dashes, though. Those haven't gone for some reason. The M dashes are here to stay. Maybe even in GPT40. All right, that's creating the code now. It's coding out. It's taking a little while, but we'll come back to that in a second. Says, "Maths is 100% insane for me. " We got Comroy from United States. We got some legends from Norway. Plenty of people from Norway here. All right, so let's keep going down. Yeah, it's a lot more advanced, isn't it, in the responses it's given for poetry as well if you just smash it out loads of poems on GPT5. Maybe you aren't. And then we've got the evaluations here, right? So GP5 is much

smarter across the board. So for maths, 94%, SWE 74%, 88% on ADA, Polyglot, blah blah. And let's compare them now. So you've got all the different modes. GT5 Pro, that's interesting. I've not seen that model yet. And then if you compare it, 03 was pretty good to be fair, but it was super slow. But if we compare this, GPT5 is absolutely crushing it. Is absolutely destroying it. Right? So this is without thinking and this is with thinking if that makes sense. So you got reasoning versus non-reasoning for competition maths. And if we have a look here, look at that. Gift 40 is all the way down there at 42% versus GT5 absolutely crushing it on maths. Let's see. Frontier map. Compare the benchmarks. It's absolutely destroying other models. Harvard maths tournament, PhD level science questions. Yeah. So, just across the board, GBT5 is crushing 03 and crushing 04 as well. Let's see. We got this. We should have this code back now by now, surely. Yeah, we got the HTML back. Let's test if that actually works. There's only one way. We can talk about it all day or we can test it, my friends. And I'm a man of action. So, let's take this HTML. We'll grab that. We'll go over to LiveWeave and we'll have a cheeky gander. So, here we go. This is a OneShot game. Let's see if it actually works. It doesn't work. Let's look at that. I'm play clicking play. It doesn't work. Let's try on something else just to make sure we haven't got this wrong. What I can do is just go over to something like claw and say preview this code and it will just turn it into an artifact that we can play. But there you didn't work. What are we doing here? I'm testing the same prompts they've got on the website. It didn't seem to work there. Maybe we need to test it on Let's see what we got here. That is the most powerful model, right? And the most expensive by a long way. It's five times more expensive than the mini version. There you go. We've tested it live. Great coding as I expected. So let's have a look here. Instruction following and aentic use 5 shows significant gains in benchmarks blah blah. Agentic searching and browsing. I do think that's going to be a big part of it. Right. But if we actually have a look here, it's 54% which is not bad, but I don't know if that's going to be good enough for using this dayto-day. What else we got here? Multimodal reasoning, health. If it can't code out a HTML game, you maybe I wouldn't use it for my health. Let's see what else we got here. faster and more efficient thinking. If you're wondering how to access all this and you just want all the links, etc. Feel free to join the AI profit boardroom. All the links are inside the SPS. Just go down to GT5 which is right at the top for you there to make it easy. Link in the comments description to that. What other questions we got? We got people from France, Luxembourg, Italy. Shout out to you legends. How can we use this for SEO? I'll come on to that in a minute. The other thing I've noticed is I've not really seen much about image generation. So, I don't think this has a new image generation model. I think it'll be using the previous one. More benchmarks here for safety. And there we go. All right. And it says on the how to use section, it says GP5 is a new default in chat GPT. Paid users can still select GPT thinking, but if we go over to chat GPT, it's not there. So, it must be rolling out later. It was only announced like an hour ago. Yeah. So, it's starting to roll out to all plus pro team and free users with access to for enterprise and edute coming in the next week. All right. I've been saying on my YouTube channel for quite a while it's going to be free. People are saying no it's not going to be free. It is free. I don't I'm not one of those guys who says like I told you. So I'm not going to say that. But yeah, you can see here you can get access for free. The difference is the usage volume, right? So for example, pro users get unlimited access to GPT5 and GPT 5 Pro will be interested to see how that performs. Plus can use it comfortably as a default model and then free users will be on like a limited tier. Right? So this is all the stuff about GPT5 coded. Some cool stuff there. All right. So let's have a look because Claude previewed it here. So we got it here. Let's test it out again. And look, it doesn't work when I preview it on Claude either. So this is interesting, right? I bet if I use that exact same prompt inside Claude Opus, it's actually going to work. So let's test that out. Right. If we go over here, let's grab this prompt like so. We'll go over to Claude, run a new chat here. We're going to select Opus 4. 1. and we'll just see, okay, can Opus 4. 1 outperform GP5 with exactly the same prompt that OpenAI has used on their website? That would be interesting to test, right? So, let's see how it goes and we'll see which one is the best for coding. Now, in the

meantime, you might be wondering, okay, like how do you what's the best way to use this for SEO? Someone asked me before, cuz it's better at coding and that sort of thing. So, let's test this out, right? What we're going to do here is I'm going to use one of my favorite content prompts. It's inside the AR profit boardroom if you want to steal this from me. We actually got a bunch of AI SEO automations inside there. So if you go to the AI SEO automation section inside the AR profit boardroom link in the comments description, you can get access to all of my best automations like you can see right here for AICO. So what I'm going to do over here is we shall go to chat GBT. We'll just use a basic model. So the chat go inside the chat here and we'll test this out. Now also you can select app inside the chat here. So if you want to get the canvas preview, you can select app. We'll do that in a second to test out. But in the meantime, what we want to do is just create an article about keyword. So let's say for example, just totally random keyword. We'll plug this in. This got all the information about me, etc. who I am. It's very contextual. And we'll use that to generate the content and we'll see how it performs as well. All right. So it's planning out the content right there. Now it's writing it. And we'll compare that side by side versus Quad Sonet in a second. That's still coding out the game. So we can't compare the outputs before. And Bare Mass says, "Julian, when do you think Google will respond to the GPT5 threat with Gemini 3. 0? " I think Gemini 3. 0 is coming very soon. I think that's going to come out. Usually what they do is they just release like as soon as everyone else does. The same with Claude. I think Claude proactively released Opus 4. 1 because they knew GP5 was coming out this week. Yeah, I think

this is something is coming very soon. Vamy says, "How do I access GP5? I'm not able to access it on the web. " It just goes to open mar if you want to get access otherwise it'll be released later today to everyone I think so it's just rolling out right rolling out slowly some people get access sooner than

others all right so this is the output from claude opus to be fair even like the already looks better than 5's output which is pretty interesting you can see here we've got the output let's see if it actually works yes it does so how interesting is that right we use exactly the same prompt from GPT5's website. G5 struggles to code it out, whereas Opus 4. 1 it out perfectly and created a really nice game out of it. If you're going to pick one model for coding, as it stands, there's only one that gives me a working output. So, that is that's really interesting. Let's see what else we got here. So, we got the content back

from here as well. What I'm going to do is I'm going to compare this output using Claude for Sonet as well, cuz that's probably the most sentient that I've seen so far for writing content, if that makes sense. So, if we go over to Claude Sonet, Claude Sonet 4 here, we'll plug in exactly the same prompts and then we'll compare the SEO content side by side and we'll just see which one performs the best, which one creates the best outputs. That's the only fair way to do it. And we're using GT5 chat here to create this. We'll see which one just feels the most human. Says, "Thanks for replying. Great work. Been following you for months. Appreciate it, too. Happy to help, mate. " And Claude is just absolutely smashing it to be honest. Let's do you know what would be interesting as well is we take the same prompt cuz I'm interested now you've got me interested. All right. So we're going

to plug in the same prompt inside Gemini 2. 5 Pro. We'll select the canvas just created directly in here and we'll see whether Gemini 2. 5 Pro can create a better output than GT5. All right. So we're going to go inside the chats here. We shall take exactly the same prompt for the jumping ball game that OpenAI recommend. Let's take that. We'll plug it into Gemini 2. 5 Pro. Plug that in and we'll just see whether that can create a working output as well. In the meantime, we've got the content back from Claude for Sonnet and we'll compare which one create the best content. All right. If we actually have a look at the prompt that we gave here, this is interesting because for example, if we have a look here, the keyword was SEO training China. And if we have a look on the outputs, let's have a look first of all at the title. The ultimate guide to dominate in Chinese search markets versus SEO training China. How to rank, win, and dominate. I would say GT5's title is nicer. Let's have a look at the intro. You're in China. You've heard SEO can work. You spent money on ads. Burned cash on agencies. Still no long-term results. Sound familiar. SEO in China is different. The rules change. Let's have a look at this content. SEO training is not just another digital marketing course. It's your ticket to cracking one of the world's most lucrative blah blah. I still prefer the content from here right from Claude. It just feels a lot more sentient. It does use the keyword as well inside the introduction which is going to help it rank better. Basic stuff but that's better with SEO. Whereas if we look at this introduction, it doesn't really mention SEO training China until quite a few lines in which is going to make it harder for search engines to rank it to be honest with you. So this is the title of the second part of the content. So why SEO training in China matters more than ever. Now, we all know, we've all seen this AI fluff word right from the beginning, right from chat GPT3, and it's still there today. So, honestly, I prefer the writing from Claude. I think it feels more sentient. It feels more natural. It's nicer to read. It's better optimized for search engines. I'm going to take the content from Claude, honestly, over Cheety5, which is a shame. I was really, really excited for cheap 5. Now, let's see how Gemini got on. If we go over to Gemini, yeah, to be fair, Gemini struggled with that. So, we've got an error straight in there. So, if we're putting them in the rankings, I would say Claude Opus 4. 1 number one, then you've got GPT5, then you've got Gemini 2. 5 Pro in this specific competition. Some people asking

like, "What do I think of generative optimization? " I'm not going to go into that because it's a little bit off topic for today, but if you want to learn about that stuff, just check out the SEO Elite Circle. SEO Elite Circle is a place to be. We've got a very active community that talks about SEO every single day. We've actually got trainings on AI SEO. Loads of cool stuff in here. So, make sure you join that if you want to learn more about SEO and actually get weekly coaching on that, too. All right.

So, I want to try Sam Alman's other prompt. Let's pull this. It's in LM Marina as well, and it is number one on the score there, which is interesting. It's apparently beating Claude Opus 4, which is interesting cuz I just I don't think it's winning at the minute, right? But I give you the real tests. Nobody's showing you this stuff, right? We're just testing real stuff live to see what actually works and what doesn't. So, let's keep going. Now, if we go over to X and let's take the prompt from SAML. So, use Beatbot to make a sick beat to celebrate GT5. All right. So, we're going to take that one. And what we'll do now is inside open router, we can compare them both, right? So, we're going to take Opus 4. 1. We're going to take GT5. We're going to make sure we use app. So, create an HTML app. And then let's hit enter on that and see what we get back from both of them. Right? So it's creating the artifacts on both. We can see which one comes back to us the fastest, which one gives us the best outputs, etc. Very easy to compare them side by side. I keep refreshing my chat GBT just to see if we've got it through, but it's not coming through. Might take a while. This is interesting.

It's actually available for free inside Cursor AI. No way. All right, let's check it out. Let's see. is that G65 is available with free credits for paying users during the launch week. All right, let's we can test that out in a sec. I'm going to sign up to Cursor AI just to show you how it performs. All right, so we got cursor open up here. It's been a while since I used it to be fair. And we got GT5 down there and we can use GT5 inside the text. So that's another way to code with it. I actually don't think I'm on a pay plan here, but I can still access GP5 and then we can switch between agent, ask, and edit. So this is really good for coding out. We'll come back to that in a second. In the meantime, we have the artifacts back. Look at this as well. So, this is interesting. E5 is still coding out, whereas Opus 4. 1 has finished already. Let's open up the artifact. We can preview this. We've got that back. Boom. Shackle. Look at that. That's looking good. That is looking good, my friend. All right, I'm going to unmute this. Yeah, it's great. It's incredible. We got big bees. Opus loves to celebrate the launch of GT5. Now, where is where's GT5 gone? There we go. All right, let's open this artifact. So, this is still coding out, right? You can see the three dots are buffering right there. So, that might take a little while. So, if you want to get access to cursor for free from what I can see, you can just code with it directly inside here. I'm not on the paid plan as far as I know. Yeah, I'm not on the pro plan. I'm on a free plan. So, if you want to get free access to GPT5 right now, you can just go to cursor. We actually have training on how to use cursor inside the air profit boardroom. Just type in cursor in the search bar and you can get access to it if you want to learn how to use this stuff and where to access it and download it etc. But yeah, you can see we've got GT5 down here. We can switch between GT5 and GT5 fast. We can use it for free as far as I'm aware. Let's see if we can do the same thing. So if we go inside the chat room, look at that 5 is still coding out. We might have to rerun the prompt, but let's take this and that's generating the output here. And there we go. All right. So that's pretty cool. So there you go. Problem solved. You can get access to it with cursor inside here. As you can see, now it looks like it's stuck inside the search on chat on open router. So, I'm just going to X off Opus. We'll try this again. Run the app, hit enter, and then we'll go from there. But as far as I'm

aware, like Opus is still outperforming it on pretty much everything. Like for speed, for the outputs, we have to rerun the prompts for GT5. I mean, for content creation called Sonic one there quite clearly. But if you want to use it for free right now and you don't have access inside chat GP just go to cursor and then you can use GD5 directly inside here and it's pretty fast inside here which is cool. The other thing that I'm noticing is cursor is a lot faster than visual studio code. Usually Visual Studio Code if I was live streaming would slow down my whole laptop whereas cursor seems to do better. Let's test some more stuff now. So if we go over to 4. 1 let's just let's take something basic like a hyperdopamine pong game. We'll go over to chat GPT GP5 and we'll use here and again we'll switch between opus and we'll compare the outputs. Right. So let's run that in the meantime. This is rinsing credits but that's the price I will pay for you my friends. That's a sacrifice is finish coding this out. Let's have a look. That's pretty cool to be fair. Wow. That was loud. To be fair that's a lot more complex than this one. If we go to here this is the output from Opus. GT5. I would say that's way more complex. We'll stop that. But yeah, that's pretty cool. To be fair, I would say that GT5 won on this round, even though we had to get it to do it twice. The outputs are really cool there. And we can switch between like humanize, swing, BPM. We can mutate the beat. Like you can see, we've got all these different options here. That is a superior output to claude opus, which the UI looks cool, but it's not as good as that. And now you can see we're getting the HTML back. Oh yeah, we're getting this GPT5 party started, my friends. Incredible. Let's take the beat down. Yeah. I would say it's pretty good to be fair. It looks like you get better outputs with open router directly when we test out the artifact. Like that's a lot more complex. But if you want to create something fun, interesting, and you want to use it for free, then just go inside cursor and just code right there. Right.

And if you need some help on that, join the AR profit boardroom. So that is it so far. Yeah, if you're wondering when chat GPT is going to be released to everyone inside the app that will come out, I would say later today. So give it 24 hours. The API is available. I've tested it today. We've tested all of the different things we can do with it, how we can build out. Honestly, for content creation, I would still stick to CL. For coding, Opus 4. 1 is definitely on the same level. And there's been certain tasks that GP can't do that Opus 4. 1 can. So, don't think that GPT is the best things. Honestly, you're going to get similar or better outputs with Opus. Anyway, we've got trainings on both. So if you want to learn how to use Opus 4. 1, we got trainings inside that profit boardroom. Link in the comments description. Otherwise, you can go over to inside the AI profit boardroom as well. We also do five coaching calls a week, so you can jump on Q& A. We have 10 master classes here. On Fridays, we do a learning master class and weekly Q& A as well. We also have a PhD scientist who specializ in AI, Dr. Jeff. So if you want to learn more about AI, this is the best place to do it. Inside the classroom, you have all these different courses. These are all the stuff that we use to grow our business. So, you can see here, for example, we've got a course for agencies. If you want to get more clients, we've got all the automations that I use for email content, social media, and videoations. If you just want to learn AI automation from scratch, we have a six week masterass with actionable homework each week to just level up and make sure you make progress. And then if you want to learn how to create a YouTube AI channel, then we have a six week road map right here that shows you exactly how I did it. Plus a bunch of cool AI SEO automations. Plus inside the community here, we just share our wins. We learn, we grow together. We have a great time doing this. success like Borat likes to say, as you can see right here. And it's a very active community, right? We're just sharing, winning, learning, growing together, and we're just constantly updating and improving things as well. So feel free to join that link in the comments description. We're currently, let's see how many members we're at. We're at 100 1,182 members right now as I'm speaking. The price does go up at 1,200 members. So, make sure you sign up now before you miss out. Otherwise, you're going to pay more later. So, make sure you sign up now. appreciate you watching and I will see you on the next one, my