**Julian Goldie** (0:00)
In this one hour course, I'm going to show you exactly how to use Kimi and Kimi K3 specifically, but this works for any other model. So if you're watching this after Kimi K3 is dropped, don't worry, this will work. It's evergreen. And you might be wondering, OK, why use Kimi? Well, basically, this is one of the most powerful, if not the most powerful model in the world right now. It's open source. It comes from China. It's cheap to use. And it's very, very powerful. I actually tested out on 50 different builds, which I'll show you in a second. And I'm going to go deep into how to use Kimi across all the different methods, because there's so many ways you can use Kimi that most people don't even realize. So it's a fantastic tool for coding. It's self-improving. It's only getting better. There's been giant leaps forward in terms of the gap between KimiK 2.7 and KimiK 3, and it's only getting better. So this is one of those powerful tools you can use, but most people don't know how to use it. And that's what we're going to get deep into inside this video. This is by far the best open source model. So I've tested GL environment 2, tested, for example, 10.7 HY3. This is the best by far. Now if you want to get started, you can go to kimi.com and we'll start using that in a second. I also want to show you that we've actually tested this out with 50 different builds using K3. So if you're like, okay, you know, what does this guy know about it? We've actually tested out. I mean, it only just dropped yesterday, but every single model of Kimi I've already tested and we've built out so many things that you can actually go and check out on Goldie Bench and it is right up there with, for example, frontier models like G56 and Fable 5 as of recording this, you can create some beautiful stuff. You can build some amazing things. I'm going to show you how we build stuff like this and how you can get the most out of it and also how you can plug Kimi into coding agents so that you can automate your work 24-7. Also be looking at KimiClaw, KimiWork, all sorts of places that 99 percent of people have never tested out when it comes to Kimi, but they're very, very powerful. So with that, let's get straight into this. We're going to start working with kimi.com. Now kimi.com is kind of like chatgbt.com, but kimi.com, you can do all sorts of agentic things. You actually can't do inside ChatGBT as well. So you'll see, for example, there's these presets like Swarm, Slides, Deep Research, Websites, Documents, and Sheets. So this is what it's going to look like once you land on kimi.com. Now you can just prompt it basically like so. So you can type something in like, for example, build me a beautiful website for an SEO agency as an example. And I just want to show you some examples of how you can use this and how you get started with this. And then what I usually like to do is just run Kimi in parallel in different tabs. So you can open it inside your browser and then run it in different tabs.
And you can see what it will start doing is it's like, right, I'm going to build this website. We're going to start using that tool. It's reading our skill and D for creating websites. And then I'm actually going to install Kimi work in the background. I think we actually just need to log in with Google if you're setting up Kimi work, but we'll come on to that in a second. So what you're going to see here is that when it starts building something like a website, it's got the chat on the left hand side and then you've got this kind of preview, this computer on the right hand side. So what this is actually doing is it's got a computer in the cloud. There's a virtual computer and then you can preview what it's doing over here. Now, if you want to minimize that, you can just click the minimize section over here, and then you can see an example of what it's doing. Now also, as it's going along, it's going to give you details of what it's working on. So for example, you don't need to be a coder to be able to use Kimi. You can just type what you want, explain it very simply, and then from there, it will start building the OWL and it will handle all the technical stuff for you. So you're at the point now where you don't need to be technical to build anything. And you can build games, apps, websites, tools, whatever you prefer using this. Now, in the meantime, what you'll see here is we've got a list of models we can use. The model that you really want to be using, if you want to get the most out of this stuff, is K3 Max. Now, you also see there's an option for K3 Swarm, which is like a team of agents, like a big, big batch of agents working together. And we'll come on to that in a second. And then we have Kimi K26. Now, K3 is the latest model. K26 is an older generation. It applies a lot faster. So if you don't want it to be like agentic, like this, so you can see it's a bit slower, it's reasoning through, it's planning, it's working in the background. If you need it to be much faster and you just want like instant responses, you don't want to wait around, then you can switch over to Kimi K26 Fast. And also you'll see here that we can switch to the context window. Now on the previous version of Kimi K26, you had 256K tokens and you had extra long, which is a million context window. You might say, what's the difference? Well, basically, there's a four times difference in the context window, which means that you can handle and process four times more. Now, if you're having a really long conversation or you're giving a lot of data to Kimi, then you'll need extra long. But if you just want like something basic like creating a website, and there's not much information that you're feeding to Kimi, and you're not planning to go back and forth with it for hours and hours, then you can just go with the basic standard context length. So we can leave that running in the background. And then what we can do over here is we can switch over to Swarm. And you'll see here that it actually gives us this option for Swarms over here.
59 more minutes of transcript below
Try it now — copy, paste, done:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090
Works with Claude, ChatGPT, Cursor, and any agent that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000777280447