Claude Code remote defaults to GLM-4.7

Anthropic's move to serve Zhipu's GLM-4.7 within Claude Code is a pragmatic pivot toward a multi-model infrastructure.

• –Leveraging GLM-4.7 allows Anthropic to scale coding-heavy remote environments without straining its own Opus 4.7 and Sonnet 4.6 capacity.
• –This represents a significant milestone: a major US AI lab integrating a flagship Chinese-developed model into its core developer experience.
• –The "hidden" default on mobile suggests Anthropic may be testing GLM's performance parity with Claude before a wider, more transparent rollout.
• –The use of an open-weights model could reignite industry debates over distillation, especially given Anthropic's history of protective stance on its own training data.