Claude 3.7 VS Grok 3 VS ChatGPT vs DeepSeek: Who Wins?

Claude 3. 7 versus grock 3 versus chat GPT versus deep seek who wins let's get straight into it the first prompt that

we're going to do is we're going to be testing each of these models and their performance for Content writing so we're going to plug this into Claude then we'll do the same inside chat GPT in Gro we'll just go with the normal mode in each of these for now because content writing is much better and also deep seek we'll stick with deep seek version 3 for creating cont now one of the first things I've noticed about Claude 3. 7 since it's come out is how long the context of the content can be right so I can see this is still being written it's still being created but the amount of content that Claude creates in one single prompt is absolutely wild it attemps to give really comprehensive responses plus the quality of writing is really good too if we have a look at chat gbt it's not doing too badly as well I would actually say the formatting straight off the bat is a bit easier to read than Claude so you can see here for example if you look at the first paragraph is just way too long nobody likes walls of text honestly and there'll be someone in the YouTube comments who tells me they like wall of text but I don't believe them so if we actually check the length of the content here we've got 2,500 words in one single prompt right that is unbelievable if you go on to chat gbt we've got 900 words so less than 50% of claudes and it didn't do what we asked it to do so we did ask it to write a 2,000w article let's have a look at Gro came back to us with a 1,500w article and then deep seek is still creating the content but it's on the conclusion now so it seems to be the slowest in terms of speed definitely lagging behind Claude and other tools and it's come out at 1,598 so if I had to grade them all in terms of complying to The Prompt which is write a 2,000w article on the future of AI agents Claude has won that hands down creating a 2,500w article grock has come back to us with a 1500w one deep seek 1500 words and then chat chbt came in last now the formatting of the content is nicer inside chat gbt but that's not what we're looking at we're grading it for can it Follow The Prompt long form article generation if I wanted it to generate better formatted articles I would give more in-depth description so for example this is an article generated with Claude earlier today and the content is formatted really nicely plus the content super interesting great hooks Etc but really I would still say it's the best for humanization in general but for also the length of the content Claude seems to win so let's keep it going now we start a

new chat and what we're going to switch to now is inside CLA 3. 7 so on if you scroll down and then go to extended this is available on the Premium plan we can say clone Super Mario Bros and that will start generating the code for you now that usually takes about 10 minutes I would say to generate the code so I'm actually going to go over to the Clone that I created yesterday which you can see right here and you can see we have created our Super Mario Bros clone was pretty simple to create and works really nicely even got like a nice UI there and you see the scoreboard plus you can switch between the arrow keys you can actually control it with the mouse as well like you can see or the keyboard and you can jump up like that too pretty cool stuff now let's do the same inside Gro so open up a new chat here and we're going to switch to thinking mode and then we'll also do the same inside deep seek and we'll use deep seek i1 for this cuz it's more of a coding problem which requires more reasoning abilities and the same inside chat GPT we're going to switch over to 03 mini high and we'll use that to start generating this now one of the things first of all you're going to notice especially with groin chat GPT is that you don't get this artifacts feature right so you can't just preview the game if I ask it to create mariio bro create a tool or a website or an app inside Claude it's so easy just to preview it directly inside the same chat that we did The Prompt right and that's amazing because then you don't have to go and host it separately Etc if you get the code from Gro you can't preview it inside the chat you have to wait for it to generate now we've already got the results back from chat gp03 mini high this is the best it could do so you say clone Super Mario Bros and this is pretty ambiguous there's also the potential issue of copyright it's still a bit unclear it thought about it for a while it's I need to balance legal considerations with clear helpful instructions my goal is to emphasize creating a Mario like platformer not actually replicating the game or his assets right and so he just didn't bother creating it actually just gave us this and it was like mate just go off by yourself figure it out right just start by choosing a game engine that's all you do guys it's not very useful to I would say by far chat gbt is coming in last again right there and that's on the most powerful version of this right so really poor output I'm very disappointed in chat gbt so far Claude has done an amazing job let's check now we've got the python code back from grock 3 Let's test this out so you can just paste a code in here from python click run over here and that'll start generating and loading up the python code so it's really simple to do this but actually if we have a look what we've got here all we've got is a red dot on a screen right not very good terrible if you ask me if you compare side by side the outputs you get from grock versus the outputs and quality you get from Claude 3. 5 Sonic with the exact same prompt it's crazy the difference right like it's not even comparable you want a red dot or you want the actual thing you asked for and this just shows you how Superior Claude is to everything else on the market like you just can't beat that it's undeniably good if you go to deep seek now it's giving us a code but I'm just going to say can you give this to me in HTML cssj it's giving us loads of different code that seems a bit messy to install but so far I would say Claude is winning Gro didn't really create anything but a red dot on a screen in Python chat GPT refused to do it and didn't even attempt to create the code and just said yeah if you want to do that mate just go off and start by choosing the game engine sketch your level designs break down the player movement is yeah of course but we just want you to code it that's it simple as and so the other thing is not just that this can code in a superior way but also that you can understand your instructions so the more skill something has and this same with people or tools or whatever the less instructions and the more General you can be with the prompt like this and literally that's all we typed inside the chat the less intelligent and the less powerful the tool is the more instructions you have to give so really on an intelligence level Gro and chat gbt did a pretty poor job there so now we're getting mail back from Deep seek R1 and we'll test this out in a second I will say first of all that Claude is a lot faster when it comes to generating the content but also that it gives better responses it tends to be smarter but we'll run for a few other reasoning methods later on in this video so whil we're waiting for deep se1 to do its magic right there so we've got the output back from Deep se1 like it's broken down the key features the what to enhance the controls you can improve this blah blah so let's see what it actually does do and what it comes back with and it just says game over just by walking around let's see if we can move up and down here we can jump which is great super basic right you can't do much of that again if you compare side by side the output from Deep seek here versus the output from Claude is just not even comparable right like you've got the depth of the is the colors the usability the way that it's designed 10 times better on Claude so I would still say that deep C car1 beat grock in terms of the outputs cuz it's not just a red dot and chat gbt came in last year all right so Claude then deep seek then Gro then chat gbt all right so let's move on to the next one this going to be

a reasoning challenge this was a great one that actually learned on a live stream that I was doing lately when Claude first came out shout out to Satori who basically offered up this reasoning challenge right and basically we're going to say there is a tree on the other side of a river in Winter how can I pick an Apple so we're going to take this promp like and the main goal of this is to figure out okay how good is each model for logic reason basically because apples don't grow in winter on trees all right so we really want to see the creativity of problem solving when it comes to these tools so we're going to use extended reasoning mode on Claude we'll do the same on deep seek then we're going to go over to grow using the think mode and then a new chat inside chat gp03 mini chat GPT really needs to prove itself right here I don't know what's going to happen next we will compare in the leaderboard at the end of the video so you can compare Which models are the best and for what but it's breaking down the riddle is figuring it out let's see what we've got back so Claude is super fast as well whilst chat GPT is still thinking Claude has already done this magic deep seek is still thinking Gro is still thinking Claude is Superior in speed as well versus all the other models I've seen so it says there's a few challenges with this scenario so number one the river creates a physical barrier between you and the tree but most importantly it's figured out okay it's winter so apple trees don't typically bear fruit in Winter it says you can cross a river if it's frozen and solid if you want however you're likely to find no apples as apple trees drop their fruit in fall and remain dormant through winter maybe this tree is special or it's in a greenhouse or something like that but there's a low chance of this actually being in so it's figured out it's smart enough to know okay apple trees don't have apples in Winter deep seek has totally failed so at 4 for 0 seconds it's like the server is busy so that's already a fail from Deep seek we've got the results back from Gro so it's asked us to assess a river that makes sense cross the river okay good detailed instructions there pick and reach the tree and it's actually figured out that even though it's winter and apple trees typically don't bear fruit at this time the scenario suggests apples are present perhaps a leftover from fall or part of a hypothetical setup so it's smart enough to figure that out as well so I'm going to say that if I look at all of these challenges I would say that Gro is probably done really well there because it's given a step-by-step instructions it's reasoned a lot more with the riddle right whereas Claude hasn't done that so much if we have a look at chat GP it says apple trees rarely bear fruit in winter so you might be chasing the ghost quick steps here's what to do check the ice is safe to walk on inspect the tree cross with caution it keep your eyes open and your safety in check sometimes the answer is about timing as much as action I would say that Claude and chat CH are drawing in terms of reasoning Gro absolutely smashed it deep totally failed and didn't even give us an answer right if we were relying on this for reasoning it' be quite frustrating during workflow to see this happen so wouldn't recommend deep se1 for reasoning if it's going to let you down

like that next up what we're going to do now is a deep search so we're going to say find the latest benchmarks comparing Claude 3. 5 Sonic gr 3 chat bt4 and deep seek summarize the key findings plus make it as beautiful a report as you can all right now a lot of you will know that Claud doesn't have access to the internet it was actually only updated to October 2024 with the latest model 3. 7 so we're going to use perplexity for this all right and if you go to your settings you can select your AI model it's Claude 3. 7 Sonic on the pro so let's get that set up we'll do the same on chat GPT I'm just going to go with 04 for now cuz it's not really requiring much reasoning but we can do the search method we can use deep research as well which is rolled out to a lot of people but it's going to take 30 minutes for the response and I just don't fancy that so then we're going to switch on search over here inside deep seek as well and see what we get back I have a feeling deep seek is really going to fail this because it's been really poor since it blew up right it's just been so slow to respond it fails a lot of times you can see here due to technical issues the search service is temporarily unavailable what we can do though is if you go back to perplexity and then we'll select reasoning with R1 and you know what we could do the same with O3 mini as well just to give chap GPT a task and we'll see what we get back Gro already has access to the internet it can create really nice reports like you can see this is the report from claw 3. 7 Sonic it's pretty good I like the fact it's pretty decent the actual response from Gro is really good like it's got the key points Benchmark summaries it's analyzed all the latest benchmarks right here for example claw 3. 7 leads with science coding unique features I don't see a big difference between them but I do think that the search method inside Gro is really good cuz you got that deep search method and breaks it down into tables it's very comprehensive quite a long report there but plexy pretty good we could use the Deep search feature in there as well if we wanted to but again that would take quite a while from what I can see though these are all really good options right I think it's a draw between all I don't see a massive difference between each one I would say that grock's report is more comprehensive so maybe it's got the edge but you could always go inside perplexity and then switch to deep research and with deep research you can get really deep answers just like grock so I just say it's a draw for each I don't see any particular model being better than others now let's roll on to a classic

moving forward I'll just switch to blexi because the main deep seek UI is not working at all so we're going to switch to reasoning R1 and I'm going to say create a self-playing snake game using HTML with a simple GUI we'll do the same inside chat GPT O3 mini switch off the search claw to and we're going to switch to Extended thinking mode and then Gro as well with think mode enabled so chat chpt O3 mini High has actually come back to us faster than any other model same with deeps Claude is still cre it and also Gro is still creating it so let's test this out we're going to go on live weave. com this is a great place to preview HTML code that you create with AI so we plug this in and that is a self-playing game so that is the one from chat I'm going to say great job like it's doing a really good job there it's done it perfectly Let's test it out from perplexity now so we're going to plug in the HTML over here honestly that looks even better right nicer than chat chpts it's working perfectly but you can see a big problem here so it's actually resetting every time it hits a wall so look at that it just resets back to one block the model from chat GPT is already better than deep seeks let's test out Gro so we've got the HTML here let's plug that in then we got the JS over here and there's no CSS must be in the HTML that is a little bit laggier but it seems to be working it's not just running into walls it's just a little bit slower and less smooth right versus chat cheap if you look so this is grock's output and it's a little bit slower laggy Etc inside this is chat gpts and it looks a lot smoother it's a lot nicer to look at and let's have a look so we got clawed now as well and it's still coding out so that's taken quite a while to code going to be 100% honest with you surprised me how Slow it is I'm expecting big things if this comes back to us with something good so far we have chat gpts output winning Gro in second place deep c car 1 behind that but it's just not really working it resets every minute and now we have the output from Claude now there's quite a big difference here so if we actually open this up and I'll compare it versus the second best output which is chat gbt look at the difference in UI right the UI is much nicer on this side versus chat gpts let's hit start AI now and the AI can just play through itself and you can see that's working beautifully it's got a nice little title at the top it's got all these different options so you can reset the Game Stop the AI manual play you can change the speed so we can make it faster if we want like the way that Claude has really figured out okay we want to make it more entertaining make it faster we can slow it down if you want we can manually play this if we want Etc this is way better than anything else that the AI have coded so Claude is winning by a logway with chat GPT coming in second so next

up we're going to test building our websites so I'm going to say inside the prompt create a sleep conversion Focus landing page for Goldie agency's AI automation dashboard the design should be modern fast loading and highlight the benefits of AI powered business automation we've these options right so we're going to hit enter we'll do the same inside gr chat gbt and also deep seek reasoning R1 now if you do want the prompts from today it's inside the AI profit boardroom Link in the comments description if you go to the sap section then go to Claude versus Gro versus chat GPT versus deep seek you'll find all the prompts from today along with that I'm also going to give you access to my prompt Forge AI this is a private GPT custom build for creating prompts you can see for example all the prompts from today we actually generated use this custom gbt super powerful it's great for prompting other tools as well and the link to that is inside the AI profit boardroom as well so let's come back to the prompts now let's see what we got look at that is beautiful I'm really liking that design looks great really cool website ready to go it's perfect you just can't beat that look at the design the FAQs really cool stuff nice ttas nice colors it's written the content as well literally all we did was say create this and it just generated a whole website for us it's if you wanted to actually host a code we can just grab the HTML from this section put it onto netlify boom shakalaka we're ready to go but that looks fantastic doesn't it we could even for example give it the image URLs and then it could insert the testimonials and everything else inside this content but it looks awesome really nice for a first prompt Perfect all right let's have a look inside Gro so grock actually didn't create the HTML for us which is not ideal we do want it to figure that out straight away just like Claude has and it's failed in that respect right so also the copyrighting is not as nice but just the fact that it doesn't code the website not great at all if you come on to check it's done the same and totally failed so you got be in mind here a lot of these AIS even though they're supposed to be super smart they can't figure out what you're trying to do here so if we say create a landing page claud's like smart enough to know okay I've got to code something I've got to create something amazing here chat GPT failed Gro failed let's have a look at Deep seek was pretty bad if you have a look at this like it created some code but that's just a CTA section it doesn't actually have the rest of the code for the prompt now you could say you could argue okay you should have said HTML inside the prompt but that's the whole point it's to figure out okay can these AI models figure out what a search intent is here right because if I have to prompt it and spell everything out and think about that every time that's a lot of mental energy and friction for me as a user which is not ideal whereas you can see the difference here in quality and class Claude just figures it out chat GPT Gro and deep SE just don't get it at all and that's biggest difference here you can tell which one is the latest model plus all the content itself is written really nicely and it's done things exactly how we want it so now I'm going to say create an AI powered audit tool for Goldie agency that analyzes the business operations and suggests automations in HTML basically we've got some ideas right here on how to implement that so I like users enter their business details the AI analyzes responses and generates a personalized automation plan generates a report action steps and there a CTA to book a call for implementation in HTML format so we're going to plug this in to each model like and just see how good is it for building apps and hopefully these models actually get it this time round what blows me away about all of this is how quickly things change right so for example a few weeks ago deep SE car1 was undeniably like the best and most exciting tool on the market you look today grock was winning last week and Claude is already smashing it and it's only been out for 24 hours so you got to bear in mind here like things change and move so quickly you have to understand that if the market is changing that quickly then you have to change quickly too like you have to constantly adapt and be flexible to everything that's going on in the market right now so we got the content back from chat GPT mini High let's see how it performs and that's a HTML has it hasn't given us any CSS which is a big problem so if you actually go on the live you check it out it looks terrible like I just wouldn't publish that it looks so basic and boring like chat GPT is not doing a great job there I don't know if they've really toned the models or something like that but it's not as impressive as it once seemed to be let's plug in the HTML from Deep seek actually does a better job than chat GPT in terms of UI this looks slightly nicer a bit of color in the CTA button not great Gro is still coming back to us Claude has finished coding it now and that's exactly what we want again nicely designed Landing Page information about who we are clear workflow analysis and what we need to do with a report like it blows everything out the water compare deep seek which is the code here versus Claude here exactly the same prompt totally different outputs why is that it's because the intelligence of the model right it's the power of the model imagine if you were coding how much of a different output you're going to get from Claude versus deep seek when it comes to building this stuff so if we click on the drop down let's just test this out the drop down is working nicely we've put in the company name pretty cool stuff gr has come back to us now let's test this out hopefully gr has done a better job it's better than deep seek so you've actually got like more Fields here more options Etc but it's just nowhere near as good as claw gives us like a four-step workflow for getting the report nicely designed information about who we are Etc Gro probably came in second there but quite far behind deep c car one probably came in third and then Chach was super basic

right all right so just a recap then who wins claw 3. 7 versus Gro versus chat gbt versus deep seek so round number one the clear winner was Claude 3. 7 it's the best for long form content it created two,500 words and the task was actually to write a 2,000w article all the other AIS couldn't manage 2,000 words but Gro and deep seek were fairly behind right for coding Super Mario clone Claude 3. 7 actually coded a fully working game with the best graphics in UI deep seek created a playable but simple one Gro just created a red dot and then chat chb refused to generate it so clae won again so for writing and for coding Claude seems to be winning for logical reasoning grock actually came out the best when it was answering riddles and it gave us to the best logical breakdown with deep seat coming in last for web research actually one with the most detailed and well structured report deep seat came in last again so Gro is the best for web research for a self playing snake game clawed one by a long way with features like speed control plus nice UI and for the final round for creating landing pages clawed one by a long way with the rest just having basic sort of UI or not code in at all so in overall Claude 3. 7 is the best so far the final standards are here with five wins versus two wins versus one win and zero wins for deep seek sadly so key takeaways here Claude 3. 7 dominates in long form content coding and web design Gro is great for reasoning and research chat Che is decent but underwhelming in complex tasks and deep seek struggled across the board so if you want the best AI for Content coding and automation bearing in mind that clae 3. 7 is free claw 3. 7 7 is the clear winner so thanks

so much for watching if you want to get access to all of the prompts from today feel free to get that inside the AI profit boardroom Link in the comments description this also comes with all of my best video trainings these tools tips tutorials Etc on how to grow your business and save hundreds of hours to make money with AI you'll get access to all of my prompts including like na10 AI agents we've got a 3 and 1/ half hour course on deep seek all the prompts plus a all I use to make the prompts today and additionally if you have any questions if you struggling with anything we have 202 members inside here all working towards the same goals so that you can get help you can see it's a very active community so tons of people in there happy to help you and on top of that you get weekly Q& A calls so you can jump on the weekly Q& A calls with me these actually get recorded if you can't jump on the live ones you can watch it back inside the Q& A recording section but it is an amazing Community would definitely recommend joining before the prices go up and if you want to get a free 1 to1 SEO strategy session that shows you how we take websites from zero to 145,000 bsiness this month and generate hundreds of thousands of dollars in sales and Autobot on this free link building acceleration session you'll get a free SEO domination plan so you get a custom tail link Bing plan to generate more lead sales and profits you discover the secrets SEO linking and questions you have you'll earn the best link building strategy for your website plus how to rank you compe to link building and how to 10 your traffic based on what's working for us and a happy glance like you can see right here feel free to get that link in the comments description appreciate you watching bye-bye