GLM-5.3

智谱最新旗舰模型，以极致后训练 Scaling 实现能力跃迁。在与 GLM-5.2 完全相同的基座上，依托数十倍规模的长程任务环境、更丰富多样的环境类型与超长周期的后训练，编程体感较前代提升 50%，并在 Terminal Bench 3.0 等公开基准中位列开源模型第一；同时涌现出强大的网络安全能力。

GLM-5.2

GLM-5.2 is Z.ai's flagship model built for long-horizon tasks, featuring a usable 1M-token context window and 744B MoE parameters (40B activated). It is optimized for project-scale engineering and agentic coding scenarios, with support for multiple reasoning effort levels, function calling, structured output (JSON mode), and streaming.