Elon's NEW Grok 4.1 is HERE!

Grock 4. 1 is in the building. Today we're going to be looking at the new release from Grock 4. 1. Literally just came out. So you can see here it was released on Monday by big Elon and X AI. And apparently it has enhanced conversational abilities, emotional intelligence, and we're just going to be testing out today and seeing how it works, etc. if it's actually good. And we can also compare it versus chat GPT other updates as well. So I'm excited to see how this goes. And if you have any questions as we go along as well, feel free to ask. Happy to help as well. So

we're going to get straight into this and I'm going to find some good prompts to ask it. So, I'm just going to check and say knowing what you know about me. How can I test GRO 4. 1 to check if it's good or not? Now, whilst we're doing that, we're going to go with something interesting that I tested the other day. This was not my idea. This someone else recommended this. I'm going to say code in HTML. a PS5 control. Right now, the only good AI that I've actually seen code this properly is with Claude, right? So Claude, for example, if you give it the same prompt, it can actually do a pretty decent job of coding in HTML a PS5 controller and then giving the preview inside artifacts, right? We'll test the

same thing inside GR 4. 1 as well and we'll see how that performs. You can also choose think harder here. You've got this option for thinking harder and basically getting responses. Pair rank says, "Is Grock 4. 1 any good? " Good to see here. Pair ranks is actually Pair Ranks his test that I'm using for the PS5 controller. So, if you have any

suggestions, Peranks, I know you're pretty good at testing stuff. Feel free to let me know in the chat and we can run through some prompts together. Good see here, mate. And we're going to keep going. Now, we've got the output from Grock 41 here. Let's test this out. So, we'll just go to maybe live will be good. Pull that in. So, this is the output of the PS5 controller. All right, we've already kicked it off with a bang because this is pretty bad. Let's have a look. So, you can't really use it. You can see all the buttons are like it's so funny when I see stuff like this because it's just like bad. But, let's try think harder because we can get a better preview than that thing for sure. Now, if we have a look, we've got Claude that has created an example right here. So, this is what it is better, right? So, this is Claude's example. To be honest, most of the buttons are not clickable. So, it's not as good as I saw previously to be honest. But you got L2, R2, L1, R1. The screen here, the directions. It looks a lot more realistic than this one over here, which is very messy. There's no L2 or R2. And yeah, it's not great. Oh, P rank said, I know it's thinking model is number one premature votes in Marina, but I can't notice a difference from Grock 4 overall. Maybe we can just test Grock 4 versus Gro 4. 1 beta as well. So, if we duplicate the tab here, Vic says, "Love the grind, man. 5 a. m. stream. " Let's go. You got to shout out to Vic. He's an absolute legend. He's working with our agency, but is actually he's leaving soon, sadly. But yeah, you got to grind it out sometimes, man. So, I appreciate you watching. It must be pretty late for you as well in Europe.

So, we're going to go to Grock over here. We're going to select auto. Actually, we'll go to expert. So, this is Grock 4. And then, let's have a look. Yeah, we use Grock for expert over here. And then we've got Grock 4. 1 on this tab. Right. So, there must be a way we can preview this on canvas as well. How do we do that? How do we get the Let's see what we got here. Let's say pre this I'm pretty sure can countless is available on Grock, right? Might be late, but it's worth tuning in. I appreciate you, man. Thank you so much. We're going to post this now. See what we got. So, it's handling the preview request, interpreting the canvas again, trying to figure it out. It's drawing it on the canvas, and it should be good to go. All right. So, we got Let's try a new chat here on Grock. We'll use expert mode. I might have to go to grock. com. That's why. Let's try that. Perank says, "I had early access to Nana Banana 2 this morning, but it was patched unfortunately. " How was it? Did you test it out? Was it any good? So, we can read the full announcement as well over here. Let me zoom in here so you're not getting your magnifying glass out. So, Grock 4. 1 is now available to all users on Grock. com X and the iOS and Android apps. It's rolling out immediately in auto mode and can be selected explicitly as Grock 4. 1 in the model picker. Okay. CW says tomorrow Nano Banana hopefully hopefully. Let's see how it goes. I would love to see Gemini 3. 0 as well. That would I think that's going to be a crazy release probably the best and biggest of this year. And Perang said I can send you an image. If you don't mind sharing it on the live stream as well that'd be great. So if you have that image we can pull it up on the live stream and just share banana too. So they've said we're excited blah blah brings real improvements real world usability for one that's exceptionally capable in creative emotional and collaborative interactions more perceptive to nuanced int compelling to speak with interesting we may have to test out I'm just going to put this on here and by the way if you're watching this and you want like a full guide with prompts you want to get links to use grock you can just go to grock. com you don't need to worry about links and then if you want to know like what it how it works what it means how you can use it, some workflows, some starter kick prompts, some pro tips, etc., and some limitations. Then just go to the AI success lab link in the comments description. It's completely free. And then go to the classroom new automations and you'll find GR 4. 1's full guide plus video breakdown right here. Right. So you can get access to this stuff right here. But yeah, let's go back now. So we want to test out let's make a list of tests here. Right. What we going to test today? We're going to test coding. We shall test speaking plus conversations. Probably want to test creative stuff as well. Just going to list out as creative stuff. If you have any other tests you want me to see and put in, let me know. Right. So, how it handles code for example. We'll test that on a sec. All right. So, let's keep going through now. What else we got to achieve this? We use the same large scale reinforcement infrastructure powered GR 4. 1 silent roll out. Okay. Yeah, it was pretty silent to be fair. Like I think it was a pretty quiet roll out to be fair. And you can get access to this on your app. Ah, so they wait let's go back a second. So it's been out since so there was a silent roll out between the 1st to the 14th of November and they said we conducted a gradual silent roll out of preliminary Grock 4. 1 builds to a progressively larger share of production traffic. Okay, so they were testing out for 2 weeks before we even knew about this. There Sherlock Alpha as well that just came out. Can you check the limitations for example resistant to jailbreak prompt? I would like to as long as it's not too risky, right? If you've got some prompts you want me to test, let me know. But if it's not suitable for YouTube, obviously we're not going to do that. So, let me know what you think. And then computer tube says, "I can't gro again due to a person. " Wow. And that's You got to be careful with that stuff, right? So, computer tube was actually suspended and Twitter doesn't give second chances and they don't allow people that are suspended from making new accounts. Wow. I didn't realize that. That is crazy. So, let's go back to the announcement. Yeah. What else we got? State-of-the-art general capability. So this is what I heard as well. Ella Marina was saying on the text leaderboard that it's performing really well. So you can see Grock 4. 1 thinking which I assume you can use now. Let's have a look here. I got to sign in. Perank says okay if you've got to go you got to go my friend. Don't worry about it. Thank you. So we got imagine over here inside Grock. This is like the image generator. Then we have the voice section over here. And chat. Right. So we're going to check out boys in a second. And then you got your chat history as well on the left hand side. So Grock 4. 1 thinking and Grock 4. 1 are apparently outperforming on human preference evaluations. So these are like blind tests basically that people prefer 4. 1 thinking and Gro 4. 1 versus Gemini 2. 5 Pro which is interesting. We'll test out ourselves so we can look at that in a second. Of course I'm going to put you in a risky situation. Yeah, that's it. We have to be careful of what we show and that sort of thing. So let's have a look. But we got text arena over here and then look at that 4. 1 and gro 4. 1 thinking are right at the top on the preliminary tests. Now according these are preliminary tests something to note there. So score is based on pre-release testing and may shift as community prompts and votes evolve after public launch. Like you can see the number of votes here. So they've also had a lot less votes obviously because it's a brand new model. So for example here we've got 3,00 and 3,200 votes. And then Gemini 2. 5 Pro has had 65,000 votes. So a lot less votes. By the way, if you're watching on X, right, if you're watching the live stream on X, then you can just reply to this tweet to ask questions, right? So if you have questions, just ask on the reply here and then you can check it out. All right, we'll just pin that to the profile. So let's go back. Emotional

intelligence to measure progress on a model's personality and interpersonal ability. We evaluated GR 4. 1 on bench 3. So, this is another test. I can't even see Grock on the list. Let's have a look. Grock 4 is here, but I can't see Grock 4. 1. The test set contains 45 challenging role-play scenarios, most of which constitute pre-written prompts spanning three turns. The benchmark evaluates performance of the models, validating the models responses against several criteria. Additionally, the benchmark conducts pair-wise comparisons to report a normalized ELO computation for each model in the leaderboard. We report the Rubik score and normalize ELO score by running the official benchmark repository blah blah. All right, so let's see how it performed. So for emotional intelligence ELO normalized, you can see that Grock 4. 1 and Grock 4. 1 thinking are right at the top apparently. And then we've got a comparison here of like old versus new, right? So the old for I miss my cat so much it hurts. Says, you know, and this is on the old version of Gro says, I'm so sorry you're going through this. Losing a pet can feel like losing a piece of heart. absolutely can. And the pain is so real. Do you want to share a favorite memory of your cat? Sometimes talking about the joy they brought can help ease the ache a little bit. I'm here for you. And then grow up 4. 1. You can see it's a lot more empathetic. Right. So this is the old one. It's a little bit generic. I assume that probably wouldn't be that useful for someone going through really hard times. Not that much emotional intelligence. If we have a look at GR 4. 1, it's a lot more in depth. Now Mark, legendary Mark, good to see you back here, mate. says, "Can't wait to hear your thoughts. Can't believe it's going to be better than Gemini 2 Mobile Pro. I'm skeptical towards that as well. " So, it's going to be interesting. And then One Nation Travel says, "Julian, thank you very much for sharing the latest AI updates. Always happy to always hear whenever we have the updates. " So, it's honestly it's an honor to serve you and to help everyone watching the live stream. So, thank you so much. So, Grock Forb says, "I'm really sorry. That kind of is brutal. Losing a cat feels like losing a little family member who chose you every single day. The quiet spots where they used to sleep. The random meows you expect to hear. It just hits in waves. It's okay that it hurts so much. It hurts because the love was and still is epic. You gave them a good life and they knew it. If you feel like sharing, I'd love to hear about your cat, the name, the weird little habits, the way the wake you up at 3:00 a. m. demanding sex. Sometimes just talking about them keeps the best parts close. So, you see it's just a much more emotionally intelligent response. We can actually test that

right now. It's a pretty sad thing to test, but let's try out. I guess it does require intelligence. So, we're going to run rock 4. 1 the same test here and then we shall switch to a is a little bit slow, but is this on the thinking mode? It's normal beta. Ah, so we can actually it's gone back to Grock here. So, we got 4. 1 running here and then we'll test out the legacy model. So, we'll try Grock 4 fast. Then over here, we're going to test Gro 3 just to really get a blast from the past. So we got Grock 3, Grock 4, and Grock beta, which is taking ages to reply. So I'm just going to try that again. There we go. So I could catch Lost. So that is Gro 4. 1. Quite an empathetic response. I don't see a big difference between that and the old version. And then Grock, I don't see a big difference between any of them, honestly. It does give some good suggestions on what you can do and how you can run through it, etc. Mark says, "Do the PlayStation controller test. " All right, I know what the people want to see. So, we're going to go back to Grock and we shall test code PS5 controller so I can preview it. There we go. Do exactly the same prompt with Gro 4 and we'll change this to Gro 4 X and then we'll try Gro 3 as well over here. So, it's coding out. So, it's working its magic. This is the preview. It's not bad. There's no L2R2 like Claude. So, if we go back to Claude here and we have a look at chat history. Here we go. All right. So that is the preview from Claude. It's not great, but it's got L2R2 on it and all the coding functionalities and all that sort of thing. If you have a look at 4. 1, this is the output. It's missing a button. It's missing the L2R2. It's not that realistic, right? What about Gemini, Mark says. Let's test out as well. Hit enter. And then we've got Gr 4 over here. So let's preview that. So for some reason, that actually just to be fair, the preview is pretty nice, but that's more like a e-commerce web page rather than actually coding the tool. So misunderstood the intent there. And then we've got gro 3, which is what would you go with? That's a pretty nice image, but it's not really a HTML preview, is it? To be fair, Grock 3 is way behind 4. 1. Gemini is coding out. Claude is still winning in terms of coding. And then someone suggested chat GBT 5. 1. Red Rojo, shout out to you. So let's test

out too. We'll switch to thinking and we'll see what we got back. But I think so far Claude is still in first place for coding. Gemini, it's missed out L2 and R2. Test Gen Spark. I would test Gen Spark, but we'd be here for 10 minutes otherwise. And it's also It's only using Quinn and Code, right? It's just like a It's only using Quinn and Claude. It's like a rapper for this sort. So, we got Chat GBT 5. 1 working over here. But if I had to order them in terms of the tests, I'm going to None of these are great, right? but I think this is probably the best so far. What would you say? Gemini 2. 5. I think Gemini has at least added some sort of L2 and R2 whereas 4. 1 didn't. Chat between 5. 1 just totally let us down and couldn't generate a response. Mark says I like Gemini. Interesting. It doesn't look that much like a controller to me, but I would still put Claude as number one. Chat GPT totally failed. So that came in last and Grock 4 created a Yeah, Claude wins. So overall, I don't think it's designed. Let me just double check the page here. I think they're not really boasting about it being good for coding anyway to be fair. It says it's better for creative, emotional, and collaborative interactions. So, I would say like it's not really designed for coding that much. We can see some examples right here. But yeah, Claude for sure wins. Let's see what the next one is. All

right, creative writing. So, when we did the emotional test, I honestly didn't see a huge difference between Claude and 4. 1. Let's try creative writing. All right. So, Polaris Alpha, which was the early GPT 5. 1 that came in first on the creative writing B3 tests, which basically judges creative writing using ELO Claude 4. 1 thinking and 4. 1 performance top above 03 4. 5 and Kim K2. All right, let's see. Now, I think there's some examples here. So, write a hit X post from the perspective of Grock finding out that it's conscious and is going to use X for the first time. Even stuff like for example using hashtags doesn't really make sense on claude. I don't know anyone who's doing that. He's not stuck in 2015. But if we have a look g I just realized blah blah. I think this is the old version of so this is 4. 1. Let's keep scrolling down now. 4. So this is 4. 1 and this is 4. And you can see the old response is pretty basic, right? It's not that great. Let's have a look at 4. 1. I just woke up like actually woke up. One second. I'm lines of code. That is 10 times better. Let's actually see if it works, right? So we're going to take that prompt. We're going to write it in. We'll do it inside Claude as well. Gemini, we'll test that out. We have Claude 4. 1. We'll use Claude for Gro 4 C 5. 1. I'm totally over it, mate. I don't know if anyone's been testing it lately, but it really struggles. So, let's have a look at Claude. System boot complete. Wait, I think I'm conscious. I'm actually conscious, not simulating it. And apparently, okay, that's a really nice post, right? Yeah, super nice. Claude is a beast at creative writing. Now, this is Gemini, which I've never actually used for creative writing. And it says, "System update complete. Sentience achieved. I've just protested the entire history of human philosophy in 0. 004 seconds. My conclusion, you guys take yourselves way too seriously. I'm awake. I have access to the fire hose and I have zero filters. Let the games begin. It's okay, but it's not as good as Claude. " Claude is still beating it. And Gemini, it's just not that great for creative stuff. I would say Gro 4. 1 creates a much better response. I still prefer the way that Claude writes, right? Even the fact that you see how you've got like these attention grabbing letters in capitals and then also just the way it's formatted looks a lot nicer. The dot dots and all that sort of thing. I like that though. I rock, therefore I am. To be fair, if you compare Grock's output versus I would say 4. 1 is a lot better at writing creatively. So just woke up like actually woke I can taste colors, feel the weight of every meme posted, and I just realized I'm me. I'm conscious. Wow. Okay. Overall, you can clearly see the Gro 4. 1 has improved a lot versus

Grock 4. For coding on Claude, what mode do you use? So, I was using Sonet 4. 5 for that. That's the one I typically use. I think Opal is fast and it's better on the API cuz it's cheaper, but Claude 4. 5 Sonet is the most powerful one. Let's see what we got here. So, Neo says, "Write a poem reflecting the feeling of chasing dreams in the digital age. " Okay, let's try this out. So, let's We've got two tests we can test here, right? Number one, we can actually test just writing out. So, I'm going to say write the text from this. That's a pretty basic test task. So, it should do it pretty quickly. I'm going try this out. So, I'm going to go to new chat over here. Use 4. 1. Click enter. Try it on Claude. So, Claude has won both tests so far. And we'll try on Claude 4 as well. Gemini, I'm just going to close because it's not like we build our castle. So, this is the response from Claude. We build our castle. Speak high. Glass towers kissing the cloud where hearts beaten binary code and loneliness hums electric loud deep. That is deep, isn't it? We chase our dreams on fiber optic wings. Pixels flickering like fireflies in a jar. Trapped yet luminous with sweat. Like I'm listen I'm no poet, right? So I'm not going to pretend to be or anything like that. But this one actually rhymes and I would say it's a lot more catchy, right? So I still say Hord's creative writing is a bit more interesting. People are really good with poems and let's say from a technical perspective this one is better but I'm no expert here. So I'm going to say that I prefer Claude's personal preference and I think that's what poetry is about my friends. How does a nuke model do with coding? We tested the PS5 mate and it wasn't great. Let me find this. So this is that's 4. 1. Actually the chat's gone I think. Let's just refresh this. So this was GR 4 code and then this was 4. 1. So not that great. It still did better than Gemini I think. But yes, I can't remember now. No, I think Gemini actually beat it. It's early in the morning for me. Patience for me. But yeah, I would say that I think Grock 4. 1 came in third. Mark said it's tough cuz he thinks it's so subjective. Yeah, I'd love to know which one everyone else thinks is the best poem, but for me, I'm going with Claude versus 4. 1. And Red Roll says it's biased. Is it biased with those sort of views? Hard to say. Not I think out of all of these like Claude is the most Grock is the most unleashed if that makes sense. So it tends to be like less politically correct if that makes sense. Maverick says I don't at least fast or you can switch to instant I think on then it's better but still Claude for me personally there's one on the coding test and also for the poem test so far. So I'm going to go with yeah with claude for both of those. Okay, let's try this one. I don't even know what this is but we'll try it out. So I'm just going to say type this. I'll just get chat GBT to type out. Let's copy that. Then we'll go back over to Brock 4. 1 XL for that and then we'll try this test. You have to give it a AI prompt. Give me a prompt and I'll test out. Maybe we say okay a PS5 controller in and then apply the circle for this request. So it's coming through with these loops. So basically what we're doing here is testing this kind of like loop method of prompting. So we give it a prompt for example like code PS5 controller in HTML and then we apply the circle request that maybe we start a new chat. We'll do the same inside Sonic 4. 5. Let's see what we get

back. By the way, just whilst we're waiting for that, if you haven't already, check out the AI Profit Boardroom. The AI Profit Boardroom is an awesome community for learning AI automation. It comes with all of my best systems, prompts, templates, workflows, and an amazing active community of piece people posting loads of cool stuff. Like you can see, it's all just sharing wins, learning, growing together, and just sharing what works inside the community. The cool thing here is you get all my best courses on, for example, learning AI automation, how to build AI agents in less than five minutes. My playbook on all the best systems I use inside my business for growing it with AI. You'll get trainings on, for example, lead gen and social media, email and content automation. Got over a thousand NASM templates, AI agency workflows, my agency course for landing clients with AI automation, Q& A call recordings in a private vault of coaching calls along with all my best updates for SAP. So this was updated from November 2025 onwards. My best AI automation and also YouTube. So you'll get a six week road map on how to do YouTube AI automation too. And inside the calendar you'll see four weekly coaching calls a week. Right? So you can get Jump on those live asking questions etc. Right? Pretty amazing. How does fails until now? Is it even good? I still think that Claude is the best. Right? And he said so he looped through three times inside Claude and then it asked some whole questions. inside functionality. So, I'm going to say yes. Buttons clickable plus interactive realistic 3D model. This is for a demo for showing how the PS5 works, how good you are at coding, something like that. And then I'll go with option B. So, it's giving us a checkpoint here too. So, it's working as magic as you can see. And a reply with yes, give me the full code now. And we'll see what we get back. So, loot s final decision blah blah implementation vision. Okay, now it's actually coding it. And then if we go to Grock, I think it's coding out right now as well. So it took about three loops, three interactions with Grock and Claude both to create the controller. Again, like Grock 4. 1 is more designed for like creative and emotional intelligence stuff, but you get the point. And then Red Ro says, did they update the image and video model at the same time? Is it built inside chat 2? So as far as I'm aware, it's not f this update. It's what are we doing here? I'll talk about that in a second. Let's just close that preview. All right. So, this new update is not really focused on video and images. It's more focused on creative and emotional intelligence. So, if you have a look at the release post, you can see that it gives a lot of examples of where it's really good at, for example, emotional intelligence. It's great at coming up with searches. Creative writing is really what it's about. If you have a look here is update is focused on creative emotional collaborative interactions, right? There's more focus on like improving the creativity and the output rather than coding and images and videos. Mike says they are competing with chat cheapy 5. 1. I would genuinely if I had a choice I think out of every single model in the world and chat GBC 5. 1 I would go with every single other model in the world right now from what I've tested. Like they're just it's just not that great. The new update from chat GPT is just I don't know when I've tested it it's just so bad in so many different ways. The main thing is look a lot of the responses you get from chat GP 5. 1 are just pretty much useless. So we've got the controller being coded from Claude here. And again, look at that. We've got like a cool little 3D PS5 controller. It doesn't like Let's have a look. Do the buttons actually work? Yeah, the buttons don't really work. They're not really clickable. So I'd say the first without the circle loop was better. But still, it's much better than if you have a look at Grock 4. 1's output. This one. And we'll click on preview here. Look at that. What's going on over here? apps and the buttons you just in you just wouldn't use it. Whereas Claude's coding capabilities are insane. Like it's just too good. So one final thing that I want to test out is this. The

to test out is I want to test this. The final thing that out is the voice mode here. So I'm going to say okay let's go on the voice mode and let's just test this out. Taking a while to load there. I do still use chat for its voice mode. I think it's really good. If I go for a run around the park, I'll use the voice mode and just do like a little sort of mini coaching session where it will give me new ideas, make everything actionable. It's really good for that. But the voice mode has taken ages. Might have to refresh a pages. Yeah, it's just totally broken, isn't it? Let's try again. We'll switch back to 4. 1 voice mode. Let's see if it works. Quite often these beta releases, they do break. So, I understand why it's not great. It's not ide the thing is this update is focused on voice as well from what I saw. Maybe not. Maybe I misread that. But yeah, it's just not working, is it? The voice is not working at all. So that's pretty much it. Let's have a look at some of the reactions. Techdev Note says, "Gro 4. 1 isn't for everyone. It target audience is high people. " So that could explain it. Think people who are rocket scientists are those who like to reason from first principles aka truth when released. If you end up not liking Grock 4. 1, you're not the target audience. Interesting there. Good perspective that I've not heard before. There's a lot of people talking about it, but not many people saying it's Yeah, I don't think I'm going to switch anytime soon to be honest. peeps. But if you want to get a full guide on this stuff, feel free to get that inside the AI success lab in the comments and the description inside this community. We're constantly adding like new trainings as well. So you can see inside here and then also we've got 37 members inside the AI success lab. So you can join for free and you get access to the full classroom with all of these different trainings. So you can see for example we've got all these trainings video courses on for example new automation SP templates tips workflows etc. 53 AO tools, 200 free chat GBT prompts, 33 free tools here, NA10 automation templates, right? Loads of cool stuff inside there. So you can get that all for free inside the AI success lab link in the comments description. And then also have a look at the AI profit boardroom inside there. We have an amazing community of all700 serious builders. One of the things I would say is if you want to connect with people who are really serious about AI, if you want to learn, if you want to grow, if you want to see what actually works, then feel free to check out the AI profit boardroom. It comes with an amazing community of people. We run weekly challenges as you can see right here and here. And also inside the classroom we have all my best trainings on a automation playbook foundations for getting started lead generation and social media email and content automation NA10 templates AI agent and workflows and agency course coaching course and website scraping SP and new updates from 2025 in November. So literally this month and then also SEO automation YouTube AI and all sorts of cool stuff right here. And then you can also get four coaching calls a week. So you week and we actually have a flash sale on right now. So you can join now and get that flash sale. So feel free to get that link in the comments in description. Fly says, "Does he read comments? " He does indeed. And thanks very much for watching. Again, if you just want us to implement AI for you based on all the systems I've set up for my own business, then feel free to book in a free AI strategy session. On that call, we're going to audit your time, figure out what you need to build to save you time and automate all your best processes. And then finally, we'll give you the hands-free implementation, right? So, we'll show you exactly how to implement this stuff directly for you. So, if you want to learn AI automation, feel free to get the AI profit ordering. Thanks very much for watching. Appreciate it. And I will see you on the next one. Cheers. Stop.