According to the research report issued by Huatai Securities, the recent iterations of the Kimi model in July and the introduction of the new GLM model in the second quarter serve as further affirmation of the core developmental trajectory of domestic models, which may exhibit shorter iteration cycles. Presently, when juxtaposed with overseas closed-source models, the technological disparity of domestic models has diminished to a mere 3-4 months. While preserving a cost-effective edge, their upper capability thresholds and pricing influence are on a continuous upward trajectory. Based on Huatai Securities' assessment, the performance of the K3 model lags only marginally behind that of GPT-5.6 Sol. Meanwhile, the GLM 5.2 model, post-launch, is poised to compete with Opus 4.7. From an investment standpoint, the swift enhancement in the capabilities of domestic models is projected to solidify the positive trends within the industrial chain. In the short term, despite advancements in pricing influence, competition at the model level is escalating. However, in the long run, the steady rise in Agent penetration rates is anticipated to further broaden the potential market scope.