Poly Hotels Details Vertical LLM Path, Lightweight Model Deployment Under RMB 100,000

Zhang Ye, deputy general manager of Poly Hotels Management, disclosed to TravelDaily the R&D path and industry enablement plan for its hotel vertical large model. Poly argues general-purpose LLMs struggle to reach core hotel operating decisions due to data-security risk and prohibitive private-deployment costs. Its approach applies hotel-specific reinforcement learning on a general model base, then distills and compresses it into a lightweight small model retaining over 95%-98% of core capabilities, with deployment cost of only RMB 200,000, or even under RMB 100,000. The model is in internal testing for business analysis and is planned for external release in October or November.