Databricks makes Chinese open-source model GLM 5.2 its default coding engine after it matched Opus at lower cost

2026-07-10

Summary

Databricks has adopted the Chinese open-source model GLM 5.2 as its default coding engine after internal tests showed it matched the performance of Anthropic’s Opus 4.8 but at a significantly lower cost. This decision is based on benchmarks using real-world tasks, which demonstrated that GLM 5.2 provides a cost-effective alternative for coding tasks without compromising on quality.

Why This Matters

This move highlights a growing trend among companies to switch to more cost-efficient open-source models without sacrificing performance. As more organizations, like Coinbase and Snowflake, adopt these models, it underscores the potential for substantial savings in AI operations, challenging the dominance of traditional Western AI providers.

How You Can Use This Info

Professionals can explore integrating open-source AI models like GLM 5.2 into their workflows to reduce costs while maintaining high performance. This shift also encourages companies to reassess their AI strategy, potentially adopting a mixed-model approach to optimize quality and cost-efficiency based on task complexity.

Read the full article