Speaker 1: The new DeepSeek R1 model just dropped. This is a game changer and it's free. And what if your AI didn't just answer you, but it actually thought like a human? Step-by-step, out loud, like a genius, spilling their secrets. It sounds absolutely insane, right? DeepSeek R1 is like ChatGPT, but on steroids, open source, smarter, faster in many ways, and it's dominating benchmarks. And it literally talks you through every decision it makes. It even thinks out loud like a genius robot diary. It's so smart, it's actually scary. And today I'm gonna show you how to use it for free, plus give you over 100 powerful use cases to get the most out of it. Let's go. Breaking news from DeepSeek as of just release, DeepSeek R1, you can see the announcement, just came out a few hours ago and it is absolutely popping off. Now this is like a rival to ChatGPT's R1 model, except it's free to use and it's completely open source. We'll get onto that in a second, plus we'll be testing it and showing you some of the best prompts you can use directly with this. Now what's interesting about this is this model comes directly out of China, but it is completely open source. Its API, if you're using the API, is 96.4% cheaper, plus you can use it for free on the chat. I'll show you how. And if you look at the benchmarks, you can see this tweet by Olleh, it's a Chinese versus American AI race. And look at these benchmarks, right? So the new version from DeepSeek R1 versus OpenAI O1 model. DeepSeek is actually beating the O1 model on many benchmarks, as you can see. So let's kick it off. The first place you're going to start. Now, if you want to get access to this, just go to deepseek.com, and then you'll see this announcement at the top. So it says, DeepSeek R1 is now live and open source, rivaling OpenAI's model available on web app and API. So you can actually get an app for it as well. And then if we click on the start now option, from here inside the chat, just make sure you click on DeepThink. This is how you're going to get access to the reasoning model. Otherwise it will just be on DeepSeek version three. So from here, we're going to say something like, what model are you? Just to double check. And you can see it's also got search integrated. But as you see here, it says, I'm DeepSeek R1, an AI assistant, created exclusively by the Chinese company DeepSeek. I'll do my best to help you. Now, this model is actually really good at reasoning and logic, obviously. So it's great for coding, building tools, and answering questions that require a lot of thinking. For example, if we put in a test like, name an English adjective of Latin origin that begins and ends with the same letter, has 11 letters in total, and for which all vowels in the word are ordered alphabetically, right? Not an easy task, sir. So we're going to hit enter, and you can see it's now thinking. So you actually get the internal thought process of DeepSeek R1. As it's going through the motions, it's like, right, okay, let's tackle the problems. The user wants an English blah, blah, blah. That's a bit tricky, but let's break it down step by step. What's very interesting about this is it actually feels quite sentient when you read the thinking process. So for example, it's breaking down exactly what it's thinking, how it's thinking, how it's breaking it all down, the thought process, step by step, and also what it's trying that's not working, right? So for example, wait, maybe sacrilegious, ends with S, starts with S, so let's check. Wait, that's 12 letters. No, it can't be that. And it literally breaks down a step-by-step process. For example, it's still thinking now, but if you actually count the word count on there, it's running through a lot of ideas. It's already done 2,000 words of thinking that is relayed to us. Super interesting, and it feels weird, like, reading it. It feels like you're looking at a human's inner monologue, and it's just going and going. So we're gonna test this out. We'll put in, for example, how many R's are in strawberry, and you can see here, like, the way that it thinks is super human, right? So it reads like an inner human monologue. For example, okay, let me start by writing down the words so I can check the letter one by one. Let me go through each letter at step. Wait, in the word strawberry, the letters are strawberry, so the R's are at position three, eight, and nine. That's three R's. Wait, that can't be right. Let me count again, and let's write the letters with their positions. Then it starts going through the motions here and highlights the three instances of strawberry. Whereas, for example, if you're using, like, a really basic model, like chat GPT-4, which is not a logic and reasoning model, and we ask it how many R's are in strawberry. So it passes the strawberry test. So now we're gonna test it again. I'm gonna say create a free SEO cost calculator, and the thing about this is you can actually run HTML directly inside the chat, right? So once it builds a tool and codes it out, you can actually run it. So it's got an artifacts feature directly in the chat here. What's super interesting is just looking through the thought process and seeing, okay, how has it built it? Why has it built it like that? And it's looking at, okay, here's the main considerations, using my ones presented on the websites, maybe suggest embedding the Google Sheet, et cetera. And then it gives us multiple options. So for example, option one creates a Google Sheet with the tool, and you can build it like that. The other option is that you can host a HTML and is self-hosted with the HTML code. And if we actually run the code like so, you can see it's working here and we can just test it out before we embed it. So for example, monthly budget, 100, keywords to target, basic services needed, hit calculate, and it does its magic perfectly. So it can build tools in seconds for any sort of website, create the HTML, and you're ready to go. And that was super fast. The other thing to note here is that it's open source, but you can use it commercially as well. Steps number one, we'll say build a Space Invaders game, and it already has knowledge of Space Invaders, plus it knows how to build it, plus it's creating it as self-contained inside the DeepSeek chat here so that we can just run it in a second. And the other thing I would say here is it seems to be really fast. When it's replying, it seems to work like really quickly. If we hit run HTML, you see it working, and yeah, it seems to be pretty crazy how easy that is to set up. Let's test something else now. So what we're gonna do from here is we're gonna grab this course outline. So we're gonna grab that semantic outline from Phrase.io, and we're gonna optimize our article perfectly based on the recommended headings, LSIs, et cetera, using Phrase, right? So Phrase actually reverse engineers the first page of Google, figures out how to optimize your article, and then gives you the recommendations. So for example, you got the recommended word count, which is 1,560 words. The keyword is best SEO trainer. And then from here, we're just gonna take these LSI and NLP keywords that are basically like semantics to optimize inside the article. So let's paste that into our prompt. And then finally, we'll add the recommended outline from Phrase as well, right? Basically what we're trying to do with the advanced reasoning model inside DeepSeek R1 is create like a perfectly optimized article based on the recommendations from Phrase, right? So we're gonna say, create the best SEO optimized article you can using the suggested headings, titles, LSIs, keywords, NLPs, and semantics in the suggested quantity below, right? It's a super logical but detailed outline prompt to see how well this article can optimize for the algorithm, right? Now, the thing to note here is it's not gonna write the content perfectly for the user, but it is gonna write it perfectly for semantics. And now it's building out the article using the semantic and NLP keywords we talked about earlier. And we'll see how well that performs step-by-step. It's really fast in just creating the content as well, which is super interesting. And let's check the word counts. We asked it to build out an article with 1,560 words. It's done 782, so that's not a great start to be honest with you. But let's plug this in the article and just see how it performs. So if we go to optimize here, yeah, you can see, look at that, 84% optimization, right? And your competitors are average of 38. So it basically followed the instructions aside from the word count almost perfectly, right? So it's included all the semantics that we asked it to with the relevant word count in the relevant quantities, like you can see right here, pretty much everything's green there, pretty wild. And it's very good at just following like detailed logical recommendations. The other interesting thing here is if you asked a human to do this, how long would it take them and how much work would that involve? A lot, right? Whereas you can just automate it using this process. And then of course, if you're gonna publish this blog for SEO, you would add the final finishing quality touches before you publish. But yeah, very impressive how powerful this is considering this is all free, right? For example, chatcheapity01, you would be paying for. This model is completely free. What we're gonna test next is creating a one-page website. So I'm gonna say create a one-page website for the niche, insert my niche here. You can customize it however you want. I wanna say make it beautiful plus modern plus sleek with relevant emojis and headings, et cetera. Keyword equals this, source context equals this and personal information about me and then link throughout the website, et cetera, to my funnels. So we'll grab this prompt, like we're gonna go back into DeepSeek R1, plug that in and you can see how it plans out. So let's have a look what it's thinking about here. So I need to create a new one-page website for the keyword. User wants it to be beautiful, modern, sleek with emojis and headings. Here's how I'm gonna start. So I'm gonna outline the structure, add the features. Styling-wise, here's what it should look like. And it's built out already. How quick was that? So it actually looks really nice little setup here. It's got, if we click on that, it actually takes us to the relevant landing page that we requested. So that works perfect. And the design is super nice, right? It's also responsive too. It's even created, for example, a name for the brand, laid out the features, et cetera, on the page. Super easy and fast to do. So you can build out websites, tools, apps, et cetera, directly inside DeepSeek R1 and it is completely free to use inside the chat. Pretty impressive stuff. Let's test it out now for building apps. So I'm gonna say create a synth keyboard that's retro plus plays music. So I am impressed with the speed. I do think that used to take ages when I used to DeepSeek's model and it's built out the keyboard with using the web audio API. So it knows which API to use, which is pretty nice too. We haven't even tested out the search model as well. I'm interested to see what that does. But yeah, we've got the retro keyboard ready to go. That is working perfectly as well. I've never seen a design like that when I've tested out similar prompts. So that looks pretty cool. And then let's just test out the search process here. You can't use DeepThink and search at the same time, actually. So for example, if you click on search, it will deselect DeepThink. So you can only use one or the other. Now, what would also be interesting is comparing these models side by side. So if we take, for example, chat GPT-01, and we compare that versus DeepSeek. So let's do that right now. And we can see which one is fastest plus which one provides the best results. So we're gonna say, create a landing page website for my SEO agency. Details below. Make it as amazing and beautiful and modern as possible. Details on my service below. We'll just grab all the information that I've already got on my website. Let's grab that. We'll plug that in, right. So it personalizes the website design. We'll paste exactly the same thing into chat GPT-01 versus DeepSeek, and we'll compare them side by side. So let's hit enter on both of those. And we'll see what we get back and how they perform. You get a lot more notes and details from, for example, DeepSeek. So if you look at the output here, you've got a lot more information from DeepSeek than you have directly from chat GPT-01. Now let's pull this up directly. So what's interesting here is, if you go into chat GPT-01, it's not actually built out the website, right? So it's not coding it, which is what we wanted it to do. Whereas if you go inside DeepSeek R1, actually created the HTML and knew exactly what we wanted, it basically understood the search intent and the intent of myself much better than R1. Now, here's the thing you have to know. R1 is a paid tool. DeepSeek R1 is free, right? But we've got a much better output from DeepSeek R1 than we have from chat GPT. So if we hit run HTML as well, it's created the website exactly how we wanted it. It's taken similar content from our landing page, even added a nice little contact page. It's super simple, but it does the job. And considering the chat GPT-01 couldn't even figure out what we wanted here, I would say DeepSeek R1 is winning in that respect. So thanks so much for watching. As you can see, DeepSeek R1 is super powerful, plus it's free and is very comparable to GPT-01. In fact, we actually got better outputs from DeepSeek version GPT-01 today. And you can see what I've actually done is created a hundred different use cases and prompts that you can use to generate whatever you want using DeepSeek and you can test those out along with a step-by-step recipe on exactly how to use DeepSeek so that you can get free access to it and understand exactly how to use it step-by-step. So if you want all of that, feel free to get it inside my free SEO course. Log into my free SEO course link in the comments description, then go to free SEO tools and then DeepSeek R1. This SEO course is absolutely amazing. It's about two or three years of work at this point with 221 lessons. And this is all completely free for you to access. It's got tons of SAPs, tutorials, tips, templates, et cetera. On top of that, what you also find in there is a section on AI agents. 2025 is the year of AI agents. So you definitely want to learn and get ahead on that. And I'll show you exactly how to build those out. If you get access to the free SEO course, you also get access to a free community of 3,400 members, a free SEO course, 53 SEO tools, and 200 free chat UBT prompts. Feel free to get that link in the comments description. And if you want to get a free one-to-one SEO strategy session, feel free to put the link in the comments description. We'll show you how we take websites from zero to 145,000 business month and generate hundreds of thousands of dollars in sales on autopilot. On this free link building acceleration session, you'll get a free SEO domination plan to discover the secrets of SEO link building. We'll answer any questions you have one-to-one. You'll learn best link building strategy for your website, plus how to quickly outrank your competitors in link building, and how to turn SEO traffic based on what's working for us and our happy clients. So if you want to get more leads, traffic, sales to your website and grow your business, feel free to put that in. Thanks so much for watching. Appreciate it. Bye-bye.