Intelligence Archive
Historical Record / Total 372 Articles
February 2026
China and Macau Win Two Golds at 2026 International Winter AI Olympiad
China and Macau secured two gold medals, one silver, and three excellence awards at the 2026 International Winter Artificial Intelligence Olympiad in Ljubljana, marking their debut at the UNESCO-backed competition.
ZTE Announces TopFlow 'Live Streaming Device' and AI Pet iMoochi for MWC 2026
ZTE has officially previewed its upcoming showcase for MWC 2026, headlined by the TopFlow 'live streaming device' and the AI-powered companion pet iMoochi. The company will also unveil its 'AI Phone Pioneer', the Nubia Neo 5 GT gaming phone with a built-in fan, and a range of AI cloud computers.
Chinese AI Models Surpass US in Global Token Usage for First Time, MiniMax Leads Charge
OpenRouter data reveals Chinese AI models have overtaken US models in global token usage for the first time, with MiniMax M2.5 claiming the top spot and Chinese models dominating four of the top five positions. The surge represents a 127% increase over three weeks, fundamentally shifting the landscape of global AI adoption.
Global AI Text-to-Video Arena: Kuaishou's Kling 3.0 Series Dominates, Capturing 7 of Top 15 Spots
Kuaishou's Kling 3.0 Pro model has secured the top spot in the global text-to-video generation rankings with an Arena ELO score of 1241, according to Artificial Analysis. The model dominates the leaderboard, taking 7 of the top 15 positions, marking a shift from entertainment demos to industrial-grade visual production.
GLM-4.7-FlashAlibaba's Qwen AI Glasses Developed Under Same Team as Quark AI Glasses
Alibaba's upcoming Qwen-branded AI glasses are reportedly developed on the same technical foundation as the existing Quark AI glasses, utilizing identical algorithms and hardware. The tech giant is set to unveil the new device at MWC Barcelona, integrating its Qwen assistant for services like ordering and ride-hailing.
GLM-4.7-FlashHuawei Cloud Launches CodeArts Public Beta: A Free AI Coding Assistant Integrating Multiple LLMs
Huawei Cloud has released the public beta of 'CodeArts', its AI-powered software development platform. The free tool integrates several large language models, including GLM-5.0 and DeepSeek-V3.2, along with Huawei's proprietary models, aiming to boost developer productivity by up to 30% in token efficiency. It emphasizes code quality, security with local file processing, and native support for HarmonyOS's ArkTS language.
Moore Threads MTT S5000 Fully Adapts to Alibaba's Three New Qwen3.5 Models
Moore Threads announced that its flagship MTT S5000 GPU has successfully completed full adaptation for Alibaba's three newly open-sourced Qwen3.5 models, leveraging native MUSA C support and Triton-MUSA compatibility to deliver high-performance inference.
GLM-5DeepSeek Unveils DualPath: Novel Framework Leverages Idle Network Cards to Accelerate Agent Inference, Breaks PD Separation Bottleneck
DeepSeek, in collaboration with Peking University and Tsinghua University, has introduced a new inference framework named 'DualPath' in a pre-print paper. The framework is designed to tackle the I/O bottleneck in long-context agent reasoning by intelligently utilizing idle storage network cards on decoding engines. It introduces a secondary data loading path to optimize KV-Cache retrieval, achieving up to 1.96x throughput improvement in online services and significantly reducing first-token latency.
Chinese AI Models Dominate Global Token Usage in February, Surpassing U.S. for First Time
Chinese open-source AI models have overtaken U.S. counterparts in global token usage on the OpenRouter platform, with MiniMax, Moonshot AI, and DeepSeek leading a surge that saw Chinese models account for over 60% of top-five usage. The shift follows aggressive pre-Lunar New Year model releases by Chinese AI firms.
Huawei Joins Agentic AI Foundation as Gold Member to Drive Open Interoperability for Agentic AI Systems
Huawei has officially joined the Agentic AI Foundation (AAIF) as a Gold Member. The company made the announcement during the Linux Foundation Member Summit on February 24, 2026. The foundation, which now comprises 146 member organizations, aims to establish open protocols, tools, and best practices for agent-based AI systems. Founded in December 2025, AAIF hosts three major open-source AI projects as founding members: Anthropic's Model Context Protocol (MCP), Block's goose, and OpenAI's AGENTS.md.
DeepSeek V4 Lite Leaked: 1 Million Token Context Window and Superior Image Generation
DeepSeek is reportedly testing a new 'V4 Lite' model, codenamed 'Sealion-lite', featuring a massive 1 million token context window and native multimodal capabilities that significantly outperform its predecessor in image generation tasks.
GLM-5Alibaba Cloud's Qwen AI 'Joins' KFC App for 'Crazy Thursday' Smart Ordering
Alibaba Cloud announced that KFC has launched an AI ordering agent named 'Xiao K' powered by the Qwen large language model. Available in the KFC app and select vehicle systems, the agent supports voice ordering, complex request handling, multi-turn dialogue, and personalized recommendations based on user history.