SenseTime has built a ten-thousand-card AI inference cluster, increasing the overall token throughput by approximately five times.

CN
22 minutes ago

179508277463124.jpg

179508277766128.jpg

SenseTime Group Inc. announced the business progress for the first three quarters of 2026, during which a ten-thousand-card AI inference cluster was established. SenseTime stated that the overall token throughput increased to about five times, and the daily average token service volume of its SenseCore cloud platform rose from over 2.4 trillion in July to 6 trillion in October.

The inference cluster adopts a heterogeneous architecture

The ten-thousand-card inference cluster established by SenseTime in the third quarter adopts a heterogeneous architecture, combining different types of computing hardware to handle AI inference tasks. In terms of chip adaptation, SenseTime indicated that over 10 Chinese chip manufacturers had completed adaptations for SenseNova Loop 9B at the time of its launch, allowing existing chips to run this model.

Unified intelligent control system integrated into enterprise processes

SenseTime integrated the unified intelligent control system into clients' business processes. In content production, SenseTime's Seko combined the AI video intelligent agent and content editing functions with a distribution network covering 180 countries and regions and over 9,000 overseas channels, providing creators a channel to distribute AI-produced content.

Model division of labor covers multi-step tasks and 8K creation

SenseTime stated that SenseNova 6.8 Flash can handle text, images, 3D content, videos, and application interfaces, and can perform complex tasks involving hundreds of steps. SenseNova U1 Pro supports repeated content revisions during the creation process and can natively output 8K; Code2Video can separately control text, formulas, trajectories, and timing in the video.

This article is reproduced from unwire.pro, and the copyright belongs to the original author and original media.

免责声明:本文章仅代表作者个人观点,不代表本平台的立场和观点。本文章仅供信息分享,不构成对任何人的任何投资建议。用户与作者之间的任何争议,与本平台无关。如网页中刊载的文章或图片涉及侵权,请提供相关的权利证明和身份证明发送邮件到[email protected],本平台相关工作人员将会进行核查。

Share To
APP

X

Telegram

Facebook

Reddit

CopyLink