Snowflake CEO finds GLM-5.2 competitive with Opus 4.7 at a fraction of the cost

Snowflake benchmarking indicates that Zhipu AI’s GLM-5.2 model performs on par with Claude Opus 4.7 on coding tasks while costing significantly less per output token. Although the Chinese model utilizes more tokens to complete these tasks, its pricing structure challenges the current market dominance of Western AI providers. This performance gap suggests that companies may increasingly prioritize cost-efficiency when selecting large language models for enterprise development workflows.