火山引擎FORCE大会追踪（1）：豆包1.8/Seedance1.5Pro发布
Haitong Securities International·2025-12-21 13:32

Investment Rating - The report does not explicitly state an investment rating for the industry or specific companies involved. Core Insights - The launch of Doubao Large Model 1.8 and Seedance 1.5 Pro at the Volcengine FORCE Conference indicates significant advancements in AI capabilities, particularly in multimodal applications and audio-video synchronization [1][13] - Doubao's average daily token usage has exceeded 50 trillion, reflecting a more than 10-fold year-over-year increase, and it serves over 100 enterprise customers, indicating successful scaling in production environments [1][14] - The introduction of the "AI Savings Plan" aims to transition AI model consumption from fragmented trials to centralized procurement, reducing friction costs for enterprises [4][17] Summary by Sections Doubao Large Model 1.8 - Doubao 1.8 focuses on solving the "last mile" issue for enterprise Agent deployment, enhancing multi-tool orchestration and reliable execution under complex instructions [2][15] - The model's capabilities are designed to support high-value scenarios such as quality inspection and retail operations, directly impacting ROI considerations for enterprise clients [2][15] Seedance 1.5 Pro - Seedance 1.5 Pro offers high-fidelity audio-visual synchronization and multilingual lip-sync capabilities, addressing common challenges in AI video generation [3][16] - The "Draft Preview" mechanism introduced in Seedance 1.5 Pro significantly improves creation efficiency by approximately 65%, facilitating standardized production processes in various sectors [3][16] Enterprise Solutions - The AgentKit and HiAgent platforms are designed to streamline deployment and integration costs for enterprises, addressing challenges in permission management and system observability [4][17] - The combination of model capabilities, platform tools, and pricing mechanisms aims to lower the total cost of ownership (TCO) for enterprises, fostering customer loyalty and reducing barriers to AI deployment [4][17]