Z.ai Ships GLM-5.3 via Post-Training

Z.ai released GLM-5.3, reusing the 743B base model from GLM-5.2 with scaled post-training. Coding gains include Terminal-Bench 3.0 jumping from 4.6 to 28.3. CyberGym reached 84.5%, surpassing Mythos 5 and GPT-5.6 Sol. Weights arrive in about two weeks after safety checks.