[stingraycharles]

Available on OpenRouter as well for those who want to test it: https://openrouter.ai/z-ai/glm-4.5

I would be interested to know where the claim of the “killer combination” comes from. I would also like to know who the people behind Z.ai are — I haven’t heard of them before. Their plans seem crazy cheap compared to Anthropic, especially if their models actually perform better than Opus.

[ekidd]

> I would also like to know who the people behind Z.ai are — I haven’t heard of them before.

To be clear, Z.ai are the people who built GLM 4.5, so they're talking up their own product.

But to be fair, GLM 4.5 and GLM 4.5 Air are genuinely good coding models. GLM 4.5 Air costs about 10% of what Claude Sonnet does (when hosted on DeepInfra, at least), and it can perform simple coding tasks quite quickly. I haven't tested GLM 4.5 Air, but it seems to be popular as well.

If you can easily afford all the Claude Code tokens you want, then you'll probably get better results from Sonnet. But if you already know enough programming to work around any issues that arise, the GLM models are quite usable.

But you can't easily run GLM 4.5 Air quickly without professional workstation- or server-grade hardware (RTX 6000 Pro 96GB would be nice), at least not without a serious speed hit.

Still, it's a very interesting sign for the future of open coding models.

[esafak]

For agentic coding I found the price difference more modest due to prompt caching, which most GLM providers on Openrouter don't offer, but Anthropic does. Look at the cache read/write columns: https://openrouter.ai/z-ai/glm-4.5

[magicalhippo]

Been playing with Grok Code Fast 1 in Cline via Open Router. It supports prompt caching as far as I can tell, and it certainly is cheap. It's been quite good for the stuff I've tried. YMMV.

[turingbook]

Actually Z.ai is a spinoff of Tsinghua University and one of the first China labs open sourcing its own large models (GLM released in 2021) . https://github.com/THUDM/GLM

[throwaway314155]

It's a spinoff of the whole university?

[cyp0633]

With a little search you can find it's a laboratory within the CS department of THU. It's a fairly large lab though, not those led by just one or two professors.

[SparkyMcUnicorn]

When it comes to "real-world development scenarios" they claim to be closer to Sonnet 4.

This is the data for that claim: https://huggingface.co/datasets/zai-org/CC-Bench-trajectorie...

[vincirufus]

Well I'd call them the poor person's claude code, wouldnt compare it with Opus but very close to Sonnet and Kimi

[vincirufus]

update the title to not seem biased / hyped