Databricks selects GLM 5.2 as default coding engine
Decision Brief
What changedDatabricks benchmarks GLM 5.2 on its million-line codebase, finds performance matches Opus 4.8 at lower cost, making it the default coding engine.
Why it mattersGLM 5.2 matches Opus performance at lower cost, showing open-source models can replace expensive closed-source ones for code agents.
Who should careTeams building on model APIs
Affected stackClaudeGLM
Builder actionWorth evaluating: can Claude replace or complement your current model
Source confidenceMedium · Reliable media or first-hand reporting
Databricks benchmarked code agents on its millions-line codebase, finding Chinese open-source model GLM 5.2 on par with Anthropic's Opus 4.8 but costing $1.28 per task vs. $1.94. Databricks will make GLM 5.2 the primary coding engine. They concluded no vendor dominates; enterprises should run their own benchmarks.
Summary basis: official / RSS sourceCompiled from the source scope noted above; the original remains authoritative.
Sources
- The Decoder：AI News
- The Decoder：AI News
留言
登入后即可留言，和其他 builder 交换实测心得。
还没有留言，抢头香。