1. SenseTime Launches Revolutionary "Daily New 5o" Real-time Multimodal Platform
SenseTime has unveiled its groundbreaking "Daily New 5o" platform, featuring real-time streaming interactions across voice, text, image, and video modalities. Positioned as a direct competitor to GPT-4o, this advanced system demonstrates exceptional capabilities in accurately describing object details and understanding integrated text-image content, marking a significant milestone in China's AI development.
2. Alipay Introduces Multimodal Medical AI Model with Trillion-Parameter Vision
Alipay has launched one of China's first medical AI models based on Ant Bailing technology, featuring trillion-parameter visual recognition capabilities. This breakthrough system achieves GPT-4 level performance in both Chinese and English medical examinations, representing a major advancement in healthcare AI applications.
3. Zhizi Engine Open-Sources Awaker Multimodal Model Suite
Zhizi Engine has released three foundational Awaker models: Chatimg for dialogue, Search for retrieval, and Gen for generation, enabling seamless integration of text and image processing. This open-source initiative significantly contributes to the democratization of multimodal AI technologies.
4. Kingsoft Office Unveils WPS AI 2.0 with Specialized Professional Assistants
Kingsoft Office has introduced WPS AI 2.0, equipped with four specialized assistants for writing, reading, data analysis, and design. This comprehensive upgrade drives office automation and intelligence, enhancing productivity across various professional workflows.
5. Moore Threads Expands AI Computing Cluster to Ten-Thousand Card Scale
Moore Threads has expanded its KUAE intelligent computing cluster to ten-thousand card scale, achieving total computing power exceeding 10 EFLOPS and supporting trillion-parameter large model training. This massive infrastructure enhancement solidifies China's position in large-scale AI computing.
6. StepFun Releases Comprehensive Step Series Large Models
StepFun has launched the comprehensive Step series, including the Step-2 trillion-parameter language model, Step-1.5V multimodal model, and Step-1X image generation model. This diverse model suite addresses various AI application scenarios with specialized capabilities.
7. World's Largest Oracle Bone Script Dataset Goes Open Source
The Digital Oracle Bone Co-creation Center has open-sourced a dataset containing 10,000 oracle bone rubbings and tracings, promoting AI algorithm development for oracle bone detection, recognition, and interpretation. This initiative bridges ancient Chinese culture with modern AI technology.
8. Humanoid Robot "Qinglong" Makes Debut with Advanced Capabilities
The full-scale open-source humanoid robot "Qinglong" stands 185cm tall with 400 TOPS computing power, supporting rapid walking, obstacle avoidance, and impact resistance. This development represents significant progress in China's humanoid robotics sector.
9. Baidu Fully Open-Sources ERNIE 4.5 Series with API Access
Baidu has open-sourced the ERNIE 4.5 series comprising 10 models with API access, featuring a 47B MoE model that demonstrates excellent knowledge retention and leading multimodal reasoning performance. This comprehensive release accelerates AI democratization efforts.
10. CITIC AI Lighthouse Factory Achieves Remarkable Efficiency Gains
AI-driven steel production scheduling has improved new product development efficiency by 56.6% while reducing defect rates by 47.3%. This technology has been successfully deployed across 29 global facilities, demonstrating significant industrial AI impact.
11. Westwell Technology Deploys Smart Port Solutions in Thailand
Mixed operations of autonomous and human-driven vehicles at Laem Chabang Port have reduced costs while improving efficiency and reducing safety incidents. This deployment showcases China's AI technology export capabilities in maritime logistics.
12. China Leads Global Generative AI Patent Applications
From 2014-2023, China filed over 38,000 generative AI patents covering artistic creation, industrial design, and multiple other fields, ranking first globally. This achievement demonstrates China's leadership in AI intellectual property development.
13. Volcano Ark Empowers ByteDance Ecosystem Products
The platform based on Volcano Engine optimizes creative experiences for Feishu and CapCut, supporting efficient content generation and collaboration. This integration showcases successful AI implementation across ByteDance's product ecosystem.
See More Content about CHINA AI TOOLS