Z.ai's GLM-5.2 sits within 1% of Claude Opus 4.8 on long-horizon coding benchmarks, runs entirely on Huawei silicon, and undercuts Western frontier models by up to 82% per token.