Transcription
Recently, people have been having a lot of issues with cloud code due to rate limits, reduced context, and several other problems that are deteriorating its overall agentic performance. But recently, I've had my hands on GLM 4.5, which was released about a month ago. It's an open-source Chinese model that has already built a strong reputation for its coding performance across the board, from reasoning to code generation to debugging and even handling multi-step agentic tasks.
Its performance for an open-source model is pretty impressive, as it can go up against and even beat certain models like Claude for Sonnet in various sorts of domains, from MMLU all the way to agentic reasoning and coding tasks. It's essentially a Claude Sonnet alternative that is open-source and reputable on all fronts, where it has the same skill set as the Sonnet model but is cheaper and, in certain cases, even better. Which is why we're showcasing it once again today on this channel, 'cause we have a way for you to actually use the GLM 4.5 directly within cloud code.
There are actually two ways that you can access it through cloud code. Now, let me just give you some context. Recently, many developers were actually switching over to GLM. And this was based off of this new plan that they had released, which is $3 a month. It's currently discounted, but it's something that is a great alternative up against the Claude Sonnet 4 subscription within Claude Code because it is fixing issues like rate limits, reduced context for just the cheap price of $3 a month, where it offers up to 120 prompts every 5 hours, and you get direct access to the GLM 4.5. It's cheaper, more reliable, and it seamlessly integrates with coding tools while having the same sort of reputable performance as a Claude for Sonnet model would.
Now, this is the paid method to set it up with Claude Code. There are other options which I'll showcase later on, but essentially, you just need to have cloud code installed using the npm install command. So, open up your command prompt and simply go ahead and install cloud code within your terminal. Once you have it installed, you can navigate to your project and then open up Claude Code with the Claude command, which is super simple. Just simply paste in Claude, and it will open up Cloud Code directly within your terminal.
Before we get started with today's video, allow me to introduce today's video sponsor, Omasi AI, a no-code platform that lets any business build a voice-powered AI sales agent in just minutes. What makes Omasi special is how customizable it is. You can create your AI agent in under five minutes and then use custom rules to control how it behaves. Wanted to upsell product B after product A, capture customer details, or guide conversations your way? Well, it's all built within. For example, just listen to me talking to JB, an AI agent sales rep.
"I want to use something to basically just use as my daily scribbling pen."
"I guess Steve, listen to me very carefully. This isn't just any scribbling. This is your daily power move. You want a pen that's reliable, smooth, and ready to keep up with your hustle."
For Shopify users, Omasi just added a killer feature 'cause your AI agent can now add items directly to your customer's cart. That means it doesn't just recommend products, it actually closes the sale right inside your store. And when customers need a real person, Omasi has seamless handoff. Just add your support number, and the AI can instantly escalate the call so buyers connect with your team directly. And after you have set your custom rules, you can then have it shared with the magic link feature, where you can send your agent anywhere, to friends, social media, or other platforms, so customers can try it instantly. So if you want an AI sales agent that's customizable, smart, and ready to drive conversions, try it out today with the link in the description below.
But with that thought, let's dive right back into today's upload. So the next step is, if you're going to be using this coding light plan, you want to make sure you first configure the API key, create it, and then you can set it up by configuring the environment variable. But for the model you want to use, you don't want to use the base GLM 4.5 model. You want to be using the coding light model card, which is an option you can switch to in the back end. And essentially, this is where you will use this model plan that is available once you have subscribed to it. And this way, you can access it directly within cloud code.
And for some reason, if you do not prefer to use cloud code, you can use Open Code, which is an open-source cloud code alternative, which is an AI coding agent built within your terminal, which also has a lot of different feature sets that wouldn't be available directly within cloud code. And to get started, it is very simple. Just get the get started command over here and just paste it into your terminal, and then you can easily install Open Code onto your computer. To open it up, you can just simply use the Open Code command, and then you can get started with it. And if you're having trouble setting the model provider, you can just simply type in Open Code o login within your terminal. And this way, you can select different providers.
And if you want to use the GLM 4.5 directly within Open Code terminal agent, you can use Enthropic as the base URL and then use the GLM provider through it. Another option is using Open Router. And this way, you can get a free API which they provide for the GLM 4.5 air, which doesn't provide the same performance as the original GLM 4.5 model, but you're going to be able to access it completely for free through Open Code, which is still pretty good.
So now when I press /models, it will showcase the Open Router model provider because I just provided my API key for it, and then I can use the GLM 4.5. And then now I can start using it to generate almost anything with Open Code. And the main reason why I prefer you actually use Open Code over cloud code is because, obviously, it's something that's open source. It has a responsive, themable terminal UI, and it also has LSP enabled, which is something that is pretty big with different terminal coding agents, as it can automatically load the correct Language Server Protocol, and it is something that's going to give you smarter, language-aware code based off of whatever model you're using. You also have multi-sessions, which can start multiple agents in parallel on the same project. And there are many other small feature sets that are available with Open Code that wouldn't be available with cloud code.
And now what you can do is just simply send in any prompt, like building a time management app. And you can see that the GLM 4.5 is powering Open Code to deploy agents to execute all these different tasks, like creating the project structure, implementing all these different features, and then building it out for me. So it looks like it has finished generating all the contents for our task management app. And in a single shot, it was able to do this, guys. This is a pretty nice-looking UI, and in my opinion, looks better than what sort of generations I would get from Sonnet with this exact same prompt. You have the ability to manage all your different tasks within the main dashboard. You have a timer, calendar like we had asked for, analytics, and a settings tab, which isn't really developed yet, but overall, it got all the components right, and you can see that it did a really great job in terms of generating the base function of our front end.
Now, in terms of UI elements, just take a look at this cloud generation of a SaaS landing page. And you can see it's pretty dull. And in my opinion, I've seen better generations of a SaaS landing page with cloud code being generated. And in this case, this is a really basic, dull SaaS landing page. Whereas this is what I was able to actually generate with Open Code, powered by the GLM 4.5 model. This looks a lot better. Styling, formatting, everything. A part of this is a lot cleaner, and this is why I believe a switch over to the GLM 4.5 model is something that you should definitely consider at this point.
If you like this video and would love to support the channel, you can consider donating to my channel through the Super Thanks option below. Or you can consider joining our private Discord, where you can access multiple subscriptions to different AI tools for free on a monthly basis, plus daily AI news and exclusive content, plus a lot more. But that's basically it, guys, for today's video on GLM coder. This is something that you can directly access through cloud code or with Open Code. And this is something that I definitely recommend that you take a look at, as it is something that's discounted for a short period of time. So if you want to use this, you can definitely get started with it with the link in the description below, or you can just simply use it like me through Open Router, which I had done for today's demo.
But that's basically it, guys, for today's video. Make sure you take a look at all the links in the description below. Subscribe to the second channel. Join our newsletter. Follow me on the Discord or the Patreon. Follow me on Twitter. And lastly, make sure you guys subscribe, turn on the notification bell, like this video, and please take a look at our previous videos because there's a lot of content that you'll truly benefit from. But with that thought, guys, have an amazing day.