At the FORCE Original Power Conference held on December 18, ByteDance's Volcano Engine presented an impressive performance: as of this December, the daily token usage of the Douyin Large Model has exceeded 50 trillion, firmly ranking first in China and third globally.

image.png

During the event, Volcano Engine officially launched Douyin Large Model 1.8 and Video Generation Model Seedance 1.5pro. The Douyin 1.8 version has been deeply optimized for multimodal Agent scenarios, significantly improving visual understanding capabilities, supporting ultra-long video analysis, and reaching a global top-tier level in tool calling and complex instruction following. Seedance 1.5pro, on the other hand, achieves millisecond-level audio-visual synchronization through a native audio-video joint generation architecture, providing strong narrative support for film and advertising creation.

The president of Volcano Engine, Tan Dai, pointed out that IT architecture is shifting toward "AI Cloud Native" centered around models. To this end, Volcano Engine not only upgraded the AgentKit development platform and HiAgent intelligent agent workstation, but also innovatively introduced the industry's first "AI Cost-Saving Plan." This plan covers all large model products, and with a tiered discount model, it can reduce enterprise usage costs by up to 47%, greatly lowering the threshold for large-scale AI implementation.

image.png

With the launch of a series of new tools such as inference outsourcing services and reinforcement learning platforms, Volcano Engine is striving to make managing Agents as simple as managing a team, helping enterprises achieve comprehensive business evolution in the AI era.

Key Points:

  • 🚀 Daily token usage of the Douyin Large Model exceeds 50 trillion, with over 100 enterprises having consumed more than one trillion tokens on its platform.

  • 🤖 Launched Douyin 1.8 and Seedance 1.5pro, enhancing the execution efficiency of multimodal Agents and achieving high-precision audio-visual synchronized video generation.

  • 💰 Introduced the "AI Cost-Saving Plan" for the first time, reducing the cost of using large models by up to 47% through technological cost reduction, further accelerating the adoption of enterprise-level AI applications.