** Hello everyone, I am Old V. A very unprofessional AI non-geek. **
It has been some time since Smart Spectrum’s GLM4.5 model was released, and I have been configuring it in Claude Code for some time. The reason why I didn’t post a review article in time was because I was using it to make an application. It really smells good. Now that the application framework is finished, I have time to write something and review my feelings of using GLM4.5 in the past few days.
As the old rule, the installation steps are not wordy, and the configuration steps are in the official documentation. https://docs.z.ai/scenario-example/develop-tools/claude or: docs.bigmodel.cn/cn/guide/develop/claude is simply exactly the same as Kimi.
Here is an interesting thing to say. When I started searching for smart music, I was actually guided to z.ai/ and used my google account to log in. I created a key and put it in Claude Code to use. Because it claims to be free, I didn’t think too much about it. After a while, I accidentally clicked on Billing on the website, which showed that I had used a Duomo knife. I’m a little confused. Is it free or not? You said it was free, why do you show how much money I spent? If you say it’s not free, it will also confiscate my money. Regardless of it, the wool continued to gather, and it stopped using it when I owed it 2 pieces of Duomo Knife. It’s a pity to think about it.
Later, I searched and found out that z.ai should be for foreigners. Let’s use https://open.bigmodel.cn. Okay, open source model. After studying the documentation, we learned that both GLM-4.5 and GLM-4.5-Air can be used in Claude Code. GLM-4.5 is used by default.
Start working when configured. This time I had a ready-made design that I wanted to make, so I threw it directly. Three steps and five divides into two. Of course, it is impossible to be 100% correct at one time, and mistakes need to be corrected later. But fortunately, Claude Code + GLM4.5 is as reliable as Claude Code+ Kimi2. But the use of tokens is really amazing. It is common to change a bug and burn 1M tokens. Because I am sorry for money, I also used QWEN Code and Gemini CLI to troubleshoot mistakes. Seeking truth from facts, these two are indeed not as good as Claude Code + GLM4.5 or Claude Code+ Kimi2. This suits my idea. Gemini CLI scripts are not as good as Claude Code. Maybe Claude Code is inherently prepared for code.
However, there is one thing that needs to be reminded. Although GLM-4.5-Air can be configured in Claude Code, the effect is much worse than GLM-4.5.
Okay, if you are interested, you can try it yourself. Newly registered users also receive free tokens: Join BigModel.cn via my link for 20M tokens! Explore AGI apps with me. Link: https://www.bigmodel.cn/invite? icode=3UCbs3vJ5bJrKdeKL%2BzG%2BEjPr3uHog9F4g5tjuOUqno%3D
##Postscript I am not a professional AI assessment, and the above feelings are very subjective. But no matter what, the wisdom spectrum is quite reliable. It can be used as another option for Kimi2.
** I am Lao V, a very unprofessional creator in the AI field. Want to learn more about AI technology trends? Welcome to follow my blog “AI Bu Know” to get the latest AI information and superficial analysis! Blog link: blog.vftl.top or https://blog.vftl.site **