Yuanxiang XVERSE-65B: The First Open Source 6.5 Billion Parameter High-Performance Model in China

AMD GPU Performance Leap! Significant Stable Diffusion Model Optimization

AMD's advancements in AI are noteworthy, particularly its latest optimizations for the Stable Diffusion model. Recently, Stability AI released an ONNX-optimized version of Stable Diffusion, resulting in significantly improved performance for AMD Radeon GPUs and Ryzen integrated graphics in AI tasks, with speed increases up to 3.8 times faster. This progress narrows the gap with NVIDIA's ecosystem...

Accelerated Transformation of Banking Technology: Large Models Deepen into Core Business

As the challenges and pressures faced by the banking industry in digital transformation intensify, more and more banks are beginning to integrate large model technology into their core businesses, rather than simply relying on chatbot applications. The latest financial reports show that some major domestic banks have made significant progress in technology investment and large model applications, but also reveal a trend of differentiated investment. According to an analysis of ten major banks by Titanium Media App, including the six major state-owned banks and several joint-stock banks, six of them have seen a reduction in technology investment. For example,

Tencent Cloud Breakthrough Upgrade! Large Model Knowledge Engine First to Integrate with MCP; AI Application Development Enters a New Era

In Chengdu's sunny April, a significant breakthrough in China's AI technology development was quietly unveiled. The 2025 Tencent Global Digital Ecosystem Summit Chengdu Conference grandly opened on April 18th, with Wang Wei, Tencent Cloud Intelligent Regional Solution Director, delivering exciting news: Tencent Cloud's large model knowledge engine has become the industry's first platform to officially integrate with MCP. This technological breakthrough means developers and enterprise users will enjoy an unprecedentedly convenient experience when building AI applications. Through Tencent Cloud's large model knowledge engine, users can easily access...

Unitree Robotics' New Patent Enables Large-Scale Dance Performances, Especially Ethnic Dances

According to Qichacha APP, Hangzhou Unitree Robotics Co., Ltd. recently published a patent for "a robot and a robot control method." The abstract shows that this invention includes a robot body and a rotating performance component; the robot body is equipped with an arm for assembling the rotating performance component and an ejection motor for throwing performance props; the rotating performance component is assembled at the end of the arm, and the ejection motor provides the ejection power and kinetic energy for the outward flight of the performance props, allowing the robot to perform at least rotating and throwing dance movements, resulting in rich movements and impressive demonstrations.

Tencent Unveils Ready-to-Use Enterprise-Grade AI Application

At the grand opening of the 2025 Tencent Global Digital Ecology Conference in Chengdu, Tencent announced its latest enterprise-grade AI application – the Tencent Cloud Large Model Knowledge Engine. This innovative tool aims to provide businesses with more flexible atomic capabilities and application development models, helping them build their own knowledge management systems. The Tencent Cloud Large Model Knowledge Engine has reportedly been successfully implemented in various sectors, including finance, energy, transportation, retail, healthcare, government, education, and tourism. Numerous companies, such as Sichuan Wen...

WORLDMEM Open Source Release: Revolutionizing Long-Term Consistent World Simulation Technology

Tencent Cloud's Wang Qi: Large Models and Knowledge Bases Empower Enterprise AI Application Landing

At the recently concluded 2025 Tencent Global Digital Ecosystem Summit Chengdu Forum, Tencent Cloud Vice President Wang Qi delivered a compelling speech on how enterprises can effectively implement Artificial Intelligence (AI) applications. He highlighted that combining large models with knowledge bases is currently the optimal path for enterprises to achieve AI implementation. Wang Qi emphasized that Tencent Cloud adheres to a "core technology self-research + embracing advanced open source" multi-model strategy, a philosophy that permeates Tencent's comprehensive layout across underlying computing power, foundational large models, model development platforms, and intelligent applications. Image source note:

Alibaba Tongyi Wanxiang First and Last Frame Video Generation Model Wan2.1-FLF2V-14B Open Source

360 Zhiyu Officially Upgrades to 360 Zhiyu and Launches the Super Intelligent Assistant, Zhi Xiaoyu

According to Liao Baicheng, Product Director of 360 Digital Intelligence Group, 360 Zhiyu adopts a product architecture of "two bases + five capabilities." The AI service base is responsible for the full life cycle management of the large model, ensuring the model's professionalism; the communication connection base unifies the collaboration functions of messaging, applications, portals, meetings, and documents, constructing a super entrance for intelligent collaborative interaction.

ByteDance Doubao Open-Source Seed Agent Model UI-TARS-1.5

The ByteDance Doubao large model team announced the open-sourcing of UI-TARS-1.5, an open-source multimodal agent built on a vision-language model capable of efficiently executing various tasks in a virtual world. The model achieved state-of-the-art (SOTA) performance on seven typical GUI (Graphical User Interface) benchmark evaluations and demonstrated, for the first time, its long-term reasoning capabilities in games and interactive capabilities in open spaces. This open-source project marks a significant advancement in multimodal agent technology for GUIs.