GLM 5.2 + Claude Code is INSANE!

**Julian Goldie** (0:00)
So today, I'm going to show you how I've plugged Glm 5.2 into Claude Code and also Obsidian. And it's a really powerful setup that means we can build and automate anything we want with Glm 5.2, which is really powerful in itself, but we can actually do it inside our harness for Claude Code, which is pretty cool. And also the cool thing about this is, obviously it's a lot cheaper than using Claude Code on a subscription. So you can get the power of Claude Code's agent harness, and you can plug Glm 5.2 into it. And on top of that, what you can actually do is you can build with it, save everything inside your workspace and use this inside your agent operating system, like you can see right here. So we have Claude Code, but it's running on Glm 5.2. So you get the same agentic power that can build whole apps, websites, whatever you want. But it's a different brain underneath, which is Glm 5.2, which is a lot cheaper. Now, at this point, you might be saying, OK, that's great. But how does Glm 5.2 compare to something like Opus 4.8?
So if we take a look at the benchmarks over here, on a lot of tasks, when I tested it out personally, Glm 5.2 could actually build better things and create better outputs than Opus 4.5. So for example, here's something we actually build out. It's kind of like a full open world game. And also bear in mind that with this system, if you ever get, for example, rate limited or you run out of tokens inside Claude Code, well, you can easily switch to something like Glm 5.2 and build out some more some stuff as well. So how does this work? So it's pretty simple and easy. What you can do is you can download Ollama for free. And then once you are Ollama, make sure you're on the cloud subscription, go to models and you can set up Glm 5.2 with Claude Code here. You can also set up in one click with OpenClaw or Hermes Agent, Codex, etc.
So it's pretty easy to plug it into all of these. And then what I actually did is I got Claude to set up for us directly and create a little workspace where we can save everything that we've created. So we have Glm 5.2 with the workspace over here. But then we've also got Glm Code, which is Glm 5.2 Cloud plugged into Claude Code here. And the other cool thing about this was it links to our memory. So this is actually synced to our memory galaxy with Obsidian, which means that every time I create outputs with Glm and Claude Code, it's personalized and it's linked to all of my memories. And also anytime we use Glm 5.2, that is actually logged directly into our Obsidian Memory2, so that we can categorize and also we can pull more context in there. And I don't need to organize it myself. So if we go back into Glm Code here, if we want to just use it, we can just tell Claude Code what we want to build.
So we can say, for example, okay, build out a Pomodoro timer app just as a quick demo here, and you can see it working directly. We've got this plugged in on high effort with Ollama, and it's pretty simple and easy to get it working. Now, you might say, okay, what is Glm Code? Well, we've got Claude Code, which is the coding agent that can actually do stuff, and then you can basically point Claude Code's brain somewhere else, like for example, Glm 5.2. Now, as another example of this, we've actually got free Claude Code plugged into the system, which means that when we're creating stuff, we can also use free local models with the same harness.
We can even direct that with voice chat, and then we can see what it builds live over here, and we've got the workspace with everything we've created over here too. It's pretty cool. And then every time we create something, it's all saved in one place. And then you can see our creation right here. Pretty simple and easy. So we can build apps with this, tools, games, landing pages. Also, the other thing that I've seen recently is like sometimes, especially if you're using Claude Code in desktop, like it will just fail at doing stuff. So this way instead, if it ever fails or breaks on you, you can just switch to Glm 5.2 with Claude Code instead. And it works in three simple steps. So Ollama runs Glm 5.2 for you, you point Claude Code at it, and then Agent OS can just make it a simple button where you can switch between all your agents and you're ready to go whenever you want. So for example, we can switch to Paperclip here or the AI Agent Mastermind. We've actually got Glm 5.2 plugged into the Agent Mastermind. This is basically like a group chat as well, where our agents can talk together. And then if we need to, we've just got in one single click Claude Code with Glm 5.2 plugged in right here. And we're going to see like our recent builds, we can see what's new with this.