A new SaferAI report says Z.ai's open-weight GLM-5.2 is nearing frontier-model capability while still lacking key safety safeguards.
Continue to AI University →