GLM-4.6

Zhipu

Zhipu GLM-4.6: 200K context (expanded from 128K in 4.5), token efficiency improved by approximately 30% compared to the previous generation.