Applied Compute

Applied Compute is the post-training company named in 171: 【AI季报 26Q2】从 coding 到 RSI，强者愈强的未来？ through its cooperation with Harvey. The source says Applied Compute helped create a legal-domain model from the GLM family, turning Open Source AI Models into an enterprise-specific model route.

In the wiki, Applied Compute functions as a bridge between open base models and Enterprise Owned Models. Its importance is that post-training, evaluation, and deployment support may let enterprises capture domain-specific advantage without waiting for frontier labs to expose the exact capability or policy boundary they need.

Connections

• Harvey — legal-AI customer or partner case in the source.
• GLM 5.2 and Zhipu AI — model ecosystem referenced by the episode.
• Enterprise Owned Models and Open Source AI Models — strategic model-ownership route.
• Model Routing Cost Control and AI Commercialization Pressure — cost and business context for choosing non-frontier-lab models.