Z.ai Aims to Catch Anthropic, OpenAI in Coding With New AI Model

Z.ai has released GLM-5.3, an updated version of its open-weight model that utilizes a 743B parameter base. By applying extensive post-training across more diverse environments and longer task horizons, the company significantly improved performance on coding benchmarks like Terminal-Bench 3.0 and DeepSWE v1.1. This development signals a strategic push by Chinese firms to compete with industry leaders Anthropic and OpenAI by offering high-capability models at lower deployment costs. The release highlights a trend of utilizing post-training refinement to enhance existing architectures.