Tencent increases Hy4 model capacity tenfold due to high demand for WorkBuddy

Tencent has introduced its flagship artificial intelligence model, Hy4, which possesses visual understanding capabilities and the ability to interact with a computer. The model was launched on November 4th.

Hy4 is capable of analyzing charts and documents, reasoning based on multiple images, as well as controlling command-line tools and graphical interfaces. The training process covers applications such as Blender and CAD software, as well as tasks requiring simultaneous operation across multiple programs.

The Hy4 model has 770 billion parameters and is available in Tencent's own products as well as through third-party platforms, including OpenRouter. Soon, the company plans to publish the model weights and a dedicated system card.

The standard API pricing remains unchanged compared to the Hy4 preview: the cost is $0.834 per million input tokens and $2.501 per million output tokens. Additionally, a 50% discount is offered on Tencent Cloud TokenHub until December 31st.

Tencent has integrated Hy4 into the WorkBuddy and CodeBuddy products, providing a limited free usage quota from November 4th to November 15th. A more cost-effective version of the model, Hy4 Flash, is also planned. The company noted that development focused on office tasks, financial analysis, and game creation, with development teams assisting in training and evaluating the model based on real workflows.

Tencent Cloud announced the launch of WAND-Vega, a multimodal application model, as well as a real-time interaction service called Real-time Intelligence and an AI voice agent for its cloud contact center. These products were presented at a multimodal session during the Tencent Global Digital Ecosystem Summit on October 30, 2026.

WAND is a Tencent Cloud brand specializing in audio and video AI. It integrates the Hunyuan model and other core multimodal models, including WAND-Vega, along with over 100 media processing tools. These tools cover generation, understanding, processing, and encoding, and companies can use them via standard APIs and existing cloud products.

The updated version of WAND-Vega is based on widely used open-source models that Tencent has significantly optimized. It is capable of generating and understanding images, videos, and speech, and is tailored for use in advertising, e-commerce, short dramatic series, and comics, as well as in AI content creation tool platforms. Furthermore, the preliminary Hy Image 3.5 model, released in September, joined WAND to create visual content for posters, storyboards, and e-commerce.

According to Tencent's statements, WAND now serves over 2000 companies and produces more than 4.5 million images and 3 million seconds of video daily, allowing clients to reduce overall content production costs by approximately 20%. The company notes that creating a short dramatic series, which previously took weeks, can now go from script to final edit in just a few hours. During the 2026 World Cup, WAND provided technical services to official broadcasting platforms in 17 countries and regions. In upcoming APEC meetings in China, the system will handle media signal distribution, AI-powered multilingual subtitling, and vertical screen adaptation for mobile devices in the media center.

The Real-time Intelligence service utilizes a global low-latency network and Tencent's real-time engineering solutions, adding speech recognition, speech synthesis, and real-time rendering capabilities. It offers two types of agents: real-time dialogue agents allow users to interrupt a conversation and ask clarifying questions without interrupting the dialogue, while real-time understanding agents convert fragmented speech into structured information. Companies can integrate the entire package or individual features, such as simultaneous translation and real-time video generation. Tencent reports that this service is applied in healthcare, logistics, AI companionship, and interview training.

The Tencent Cloud AI voice agent for Contact Centers uses a dialogue model optimized for phone calls to handle incoming calls, outbound follow-up contacts, and responses exchanged with human staff to solve problems, not just answer them. It is implemented in the automotive, financial, insurance, logistics, and manufacturing industries.

EdgeOne Booster is used for inference operations, providing network acceleration, security, and model deployment on edge nodes located close to users. Tencent claims that thanks to this, the Moya robot from DroidUp became 30% faster in providing the first response and reduced interaction jitter by 4.6 times. Tencent Cloud also collaborates with AgiBot and DroidUp in embodied AI.

Tencent's Executive Vice President, Dongsong Tang, announced at the Tencent Global Digital Ecosystem Conference on October 29 that the company plans to soon release the Hy4 and Hy4 flash models, adding the function of visual understanding to its lineup of artificial intelligence models.

Compared to the preliminary version of Hy4 presented in August, the full Hy4 model will demonstrate improved performance in tasks requiring extended reasoning and execution, including real productivity workflows. Furthermore, Tencent aims to enhance reasoning efficiency.

The Hy4 flash model will be focused on cost-effectiveness while retaining visual understanding capabilities and AI agent functions. The preliminary Hy4 version has a total of 770 billion parameters, with 49 billion parameters activated during inference, and it supports a context window of one million tokens. Tang noted that Tencent is jointly developing its models and products to improve their efficiency in practical applications.