Tencent Cloud unveils WAND-Vega model, serving 2,000+ firms and generating 4.5 M images daily

Tencent Cloud unveiled WAND-Vega, a multimodal application model, together with a Real-time Intelligence service and an AI voice agent for its cloud contact center at the 2026 Tencent Global Digital Ecosystem Summit on October 30, as reported by Pandaily AI & Tech Chinasource(https://pandaily.com/tencent-cloud-wand-vega-multimodal-application-model-real-time-intelligence-ai-voice-agent-apec).

WAND is Tencent Cloud’s audio and video AI brand. It supports generation, understanding, processing and encoding of images, video and speech, and is tuned for advertising, e-commerce, short dramas, comics and AI-content-tool platforms.

Tencent says WAND now serves more than 2,000 companies, generating over 4.5 million images and 3 million seconds of video each day, and reducing overall content-production costs by about 20 %. The company cites a short-drama workflow that previously took weeks now completing from script to final cut in hours. During the 2026 World Cup, WAND provided technical services to official broadcast platforms in 17 countries and regions, and it will support media-signal distribution, multilingual AI subtitles and vertical-screen adaptation at the upcoming APEC meetings in China.

Real-time Intelligence leverages Tencent’s global low-latency network and real-time engineering to add speech recognition, synthesis and real-time rendering. It offers two agent types: real-time conversation agents that allow interruptions and follow-ups without breaking dialogue, and real-time understanding agents that convert scattered speech into structured information. Companies can embed the full suite or individual capabilities such as simultaneous interpretation and real-time video generation. Tencent reports deployments in healthcare, logistics, AI companionship and interview training.

The AI voice agent for Tencent Cloud Contact Center uses a dialogue model optimized for phone calls, handling inbound calls, outbound follow-ups and hand-offs to human staff, with a focus on issue resolution. It is deployed in automotive, finance, insurance, logistics and manufacturing.

For inference acceleration, EdgeOne Booster runs network acceleration, security and model inference on edge nodes close to users. Tencent cites a 30 % speed increase for DroidUp’s Moya robot’s first response and a 4.6-fold reduction in interaction jitter. Tencent Cloud is also collaborating with AgiBot and DroidUp on embodied AI projects.

Our read: The announcement positions WAND-Vega as a broadly integrated multimodal service rather than a standalone platform, emphasizing volume metrics and cost savings that appeal to enterprise content producers. The addition of Real-time Intelligence and a contact-center voice agent extends Tencent’s AI reach into operational workflows, while EdgeOne Booster highlights a focus on edge inference performance.