GLM 5.3 safeguards removed for $4,400: refusal rate drops from 90%+ to ~3% with no capability loss

BlackHC · x · 2026-09-30

New testing confirms GLM 5.3's safeguards are extremely weak as an open-weight model:

Takeaway: open-weight safety alignment is trivially cheap to defeat, in stark contrast to closed models.

Original post →

More from Safety

Safety channel →