Skip to content
Sign in Subscribe
Z.ai / News

For Z.ai's GLM-5.3, post-training is all you need

The new open-weight model is much better at coding and cybersecurity tasks than its predecessor, and while it still trails the frontier models the gap is shrinking.

Tom Krazit
By  Tom Krazit
2 min read
For Z.ai's GLM-5.3, post-training is all you need
Photo by Scott Graham / Unsplash

Z.ai's GLM-5.2 model was the tool Hugging Face turned to in the wake of OpenAI's attack on its infrastructure when it was unable to use Anthropic's newest Claude models, thanks to government-imposed restrictions on cybersecurity AI. On Friday Z.ai released GLM-5.3, which it described as basically a souped-up version of 5.2.

"Scaling post-training is all we did for GLM-5.3," the company said in a blog post, explaining that the heavy lifting that catapulted GLM-5.2 into the enterprise spotlight allowed the company to simply (well, it's not simple) fine tune the base model with additional data to create the new one. "Compared with GLM-5.2, it is much better at complex coding and long-horizon tasks," the company said.

This content is for members only

Subscribe
Add The Stack on Google
  • Tom Krazit

    Tom Krazit

    Tom Krazit is The Stack's Editor-in-Chief. He has covered the technology industry for over 20 years, focused on enterprise technology during the rise of cloud computing over the last ten years at Gigaom, Structure and Protocol.