Discover how to access GLM 5.3 for free using Zed Code, a powerful frontier-level AI that rivals Claude and GPT in coding benchmarks. This video covers the easy setup process, real-world performance data, and how to automate your development workflow today.
Full transcript
You can now get access to GLM 5.3 for free and it's available in Zedcode. Now if you're wondering why you would want free access to this free model, it is right up there with Fable 5, Claude Fable 5. It's frontier level technology but the difference is that with GLM 5.3 you can use it for free. Now if you're wondering how to get access, basically new users can try GLM 5.3 on the free tier, no subscription required.
All you have to do is just get Zedcode. So Zedcode is an app you can download. It looks very similar to for example something like Claude Code and then once you've done that you can download this and start using it for free. They do have upgraded plans but you can use it for free right now and you get a free trial when you sign up and you don't need to put a card in or anything like that.
So for example once you've signed up to Zedcode it's going to look something like this and you can just select your project then you can select GLM 5.3 and from here you can change your permissions and then build out whatever you want to. So for example we actually built out this website you can use it for building apps, games, tools and it can build and automate whatever you want and you can run a few tasks on it for free. Obviously you don't get that many tokens on it for free but just want to be 100% transparent with you. And then inside the left hand side here you've got search, you've got automation, so you've got skills and new tasks.
Now you might be wondering okay why would you use GLM 5.3? So for me what I actually like to do is just get Claude to orchestrate it. So for example I can actually tell Claude to build stuff with Zedcode directly and then I don't even need to manage it. I've got my AI to manage it directly for me and also if you look at the benchmarks here if you compare it for example on Terminal Bench 2.1 GLM 5.3 scores 88.2 versus Fable 5's 88.
Terminal Bench 28.3 versus Fable 5's 33.7. So they're all within like one to three points, maximum five points of the benchmarks here and actually GLM 5.3 outperforms Claude Fable 5's best model here as you can see on SWE Marathon v1.1. So it scores 42.5 versus 33.1 versus 42.5 from GPT 5.6 Sol. So if you're looking for like a free frontier level AI this is pretty good as well and it can also plug into agents as well as a CLI.
One thing to note is that it will be going open weights as well so they're going to drop the weights for it on Hugging Face soon. Bear in mind like you won't be able to host it locally unless you have some absolutely wild local setup right. They're just going to release the open weights and that makes it open source but it doesn't mean like you can run it locally or anything like that and what it's ideal for really is coding. That's one of the best things it's best at so definitely worth checking out.
By the way if you want to get more courses and trainings on stuff like this feel free to check out the AI Profit Boarding. I'll answer your questions personally inside the community, inside the classroom, you can access all of my best trainings in the calendar. You can jump on four weekly coaching calls and then also we're dropping a one-hour course on how to use GLM 5.3 very shortly so you'll get first access to that and you can meet people in your local area from the map who are building with AI agents and coding agents like this. Hope to see you inside there.
Cheers for watching. Bye bye.
More episodes