QQ Pets Reborn with an AI Brain: Integrated with Tencent Hy3 Large Model, This Time It Will Actually Respond to You

Tencent has officially announced that QQ Pets, which carry the nostalgic memories of a generation, will return in a new form. This upgrade not only fully 3D-izes the appearance and gives it a soft, tactile texture, but also embeds Tencent's self-developed Hy3 large model into the pet's little brain, giving this virtual companion the ability to respond actively.

The pet-raising gameplay retains classic mechanics such as feeding, bathing, and working, while being comprehensively upgraded - you can choose a pet with various appearances and personalities, whether a drama queen or a clingy type, each with its own character. The pets will also appear in core QQ scenarios such as the contact list and chat window, always by your side.

Interactions between friends have also been refreshed: you can visit your friend's home, feed or bathe their pet, and even step on the spark. Tencent stated that users can even raise their own pets to become top stars, experiencing a more immersive breeding experience.

29.5 billion parameter large model drives, competing directly with flagship models

The Hy3 model that brings soul to QQ Pets was officially open-sourced on July 6th, using a MoE (Mixture of Experts) architecture, with a total of 29.5 billion parameters, 21 billion activated parameters, and a maximum context length of 256K. In key tasks such as reasoning, intelligent agents, and long contexts, Hy3 achieves results comparable to flagship models with 2 to 5 times the parameter scale. Installing a large model of this scale into QQ Pets means that every time you feed or interact with your pet, there is top-tier AI power driving it behind the scenes. QQ Pets may no longer just be a nostalgic symbol of the digital age, but is becoming a clever entry point for AI to enter the daily lives of hundreds of millions of users.

WeChat Work AI Assistant Closed Beta: Write a Weekly Report in One Sentence, Tackle 99+ Group Chat Messages

Tencent's WeCom AI assistant "Da Yuan" begins beta testing. It wakes via phone edge swipe or PC sidebar, no app switching. Key feature "climbing floors" auto extracts group chat highlights, summarizes voice messages, and generates content from conversations, handling daily office tasks in one stop.....

Unlock 'Touch and Share'! Tencent WorkBuddy Makes Its Debut on the HarmonyOS PC App Market

Tencent's WorkBuddy, the first desktop office AI agent on HarmonyOS, is now available on the HarmonyOS PC app market. Integrated with AI, it overcomes login/connectivity issues, adapts system permissions, and introduces a 'tap-to-share' feature using HarmonyOS capabilities, not found on other systems, to boost office productivity.....

Tencent Yuan and Multimodal Models Merge: The Multimodal and Large Language Model Departments Combine, YAO Shunyu Oversees the Full Modal Upper Limit

According to a report on July 24th, Tencent announced on July 23rd that it would merge the Yuan Multimodal Model Department with the Large Language Model Department to form the Basic Model Department, led by Chief AI Scientist YAO Shunyu. This move aims to enhance R&D and collaboration efficiency, and strive to achieve the intelligence upper limit of full modal models. There were hints of this integration earlier, as YAO Shunyu had already taken charge of the Large Language Model team in December last year. Now, combining both sides means Tencent is concentrating resources to promote the deep integration of multimodal and language models, accelerating the construction of a new generation of unified basic models, and moving towards higher peaks of full modal intelligence.

Tencent Unveils WorkBuddy Bench: A Coding Intelligent Agent Testing Ground Integrated with Code, Web, Office, and Security

Tencent has released the WorkBuddy Bench multi-domain evaluation suite, with a paper published on arXiv. It breaks the fragmented approach to evaluating coding intelligent agents and the lack of transparency in production benchmarks, integrating four types of work scenarios — repository-level code engineering, front-end artifacts, office automation — into one platform. The biggest highlight is not the volume of questions, but rather the design of the questions themselves, which prevents memorization of answers, ensuring that the evaluation truly reflects the generalization and transfer abilities of intelligent agents across different domains.

Tencent Launches New Generation AI Creative Agent Miora, Design Beginners Can Also Complete Full Visual Solutions in One Click

Tencent's new AI creative agent Miora is now fully launched, and new registered users can get 1000 points. This agent features "memory, understanding, and collaboration" as its core characteristics. Users only need to propose the final goal, and complex tasks will be automatically broken down and completed by coordinating multiple agents. A multi-expert collaboration system covers diverse creation scenarios and actively communicates during the process, significantly improving the efficiency of intelligent creation.