Phala Network (PHA)

I wanted to share how we avoided spending roughly $160k/year to host GLM-5.2 with its full 1M context.

When GLM-5.2 launched, Phala was one of the launch partners of https://t.co/739QZk2emc. We tried to bring it up immediately on our existing 8×H200 setup. The model was exciting, but the deployment reality was less romantic: we could not fully open the 1M context window on that node. The obvious path was to move to a more expensive setup, likely Blackwell-class hardware. That was not a small cost decision.

This is where open source becomes powerful. Instead of treating the model as a fixed artifact, the team started asking whether we could make the memory budget work. They quantized the routed MoE experts to 4-bit, kept the important parts in FP8/BF16, and validated the result carefully. The outcome was GLM-5.2-W4AFP8: full 1M context on a single 8×H200 node, with benchmark results aligned with the FP8 baseline.

Until today, GLM-5.2-W4AFP8 on Hugging Face has already close to 20k downloads. I think that sa

Our LegalCiteBench paper was accepted to AI4Law @ ICML 2026 Workshop 🎉

Authors: Sijia Chen, @bgmshana, @zhou49

It studies a narrow legal AI failure mode: can LLMs produce correct case citations without retrieval?

Paper: https://t.co/ZhwjhjuuWd https://t.co/9SIQEfSpOT

Here is the technical document produced in collaboration with @oppo: https://t.co/D917G8PudH