Express News | All Tencent Buddy products now support Kimi K3 in their built-in models.
Express News | Tencent Launches WorkBuddy App with HarmonyOS Version Simultaneously Released
Tencent Cloud Launches Overseas Version of ADP 4.0
Gelonghui, July 19 — At the 2026 World Artificial Intelligence Conference yesterday, Tencent Cloud officially launched the international version of its Tencent Cloud Agent Development Platform (ADP) 4.0, along with simultaneous upgrades to three core modules: the Intelligent Workbench, Claw Mode, and Skill Plaza.
Goldman Sachs Comments on Kimi K3: Chinese Open-Source Models Reach a Critical Inflection Point in Global Adoption, with Intense Competition in High-End Models
Goldman Sachs stated that Kimi K3 has emerged with 2.8 trillion parameters, raising its API pricing to $2.30 per million tokens—the highest price yet for a Chinese AI model—marking a shift among domestic AI firms from a 'price war' to a battle for 'pricing power.' However, on the flip side, Zhipu’s stock plunged 28% in a single day and MiniMax dropped 16%, revealing deep market concerns about the sustainability of competitive moats among AI model companies amid intensifying competition in the high-end programming segment.
Express News | Tencent Upgrades and Launches "Embodied Intelligence Full-Stack Solution"
Live from WAIC | Tencent Vice President Lin Songtao: The gap in models is narrowing, but the gap in engineering is widening.
Sina Tech News, July 18 (morning): The 2026 World Artificial Intelligence Conference and the High-Level Meeting on Global Governance of Artificial Intelligence is being held in Shanghai from July 17 to 20. Today, Tencent hosted its AI Application Innovation Forum. Lin Songtao, Vice President of Tencent, stated at the event: "The gap between models is narrowing, but the engineering gap is widening. The performance differences among mainstream models in inference are actually growing wider, yet whether this can be transformed into enterprise productivity or enhanced efficiency for individuals and organizations remains a widely discussed issue." He added: "Models largely determine the upper bound of AI, while engineering determines the lower bound. Users will stay because of a good"