Z.ai's GLM-5.3 boosts coding and cybersecurity performance without adding parameters

Z.ai released GLM-5.3 in August 2026, achieving a 50% improvement in programming capabilities over its predecessor GLM-5.2 while keeping the same 743 billion parameters and base architecture unchanged. The gains came entirely from post-training refinements, including better reinforcement learning methods, improved data quality, and a new long-context processing architecture. GLM-5.3 reached the top of global cybersecurity benchmarks, including first place on CyberGym for vulnerability detection, and uncovered 2,436 vulnerabilities across 269 real-world projects, among them a DNS protocol bug dating back to 1983. Z.ai plans to open-source the model weights within two weeks, alongside a controlled-access programme and a community-driven security initiative. The results challenge the prevailing assumption in AI development that larger models and more training data are the primary drivers of performance gains.