GLM-5.3 is here with advanced cyber capabilities — and reportedly already found a 'serious vulnerability' in Cursor

GLM-5.3 is here with advanced cyber capabilities — and reportedly already found a 'serious vulnerability' in…

Z.ai released GLM-5.3 via post-training, scoring 28.3 on Terminal-Bench (6x gain) and unexpectedly advancing cyber exploitation; discovered Cursor vulnerability. For enterprise coding agents, post-training scaling proves efficient without retraining; rapid offensive progression signals governance risk.