China AI

AIArtificial Analysis finds AMD's MI355X delivers 15% to 40% more throughput per MW than Nvidia's B200 on DeepSeek V4 Pro 0813 between 35 and 75 tok/s per user, peaking at about 94M tokens/s per MW at 33 tok/s per user.

AIArtificial Analysis says Kimi K3, with 2.8T total and 104B active parameters, serves about 8M tokens/s per MW on AMD MI355X at 50 tok/s per user, roughly a third of the roughly 24M reached on NVIDIA B300 with multi-node disaggregated configs.

AIArtificial Analysis says AMD's MI355X serves about 26M tokens/s per MW at 100 tok/s per user on GLM-5.2/5.3.

AIArtificial Analysis adds AMD's MI355X to its AA-AgentPerf agentic inference benchmark for Kimi K3, GLM-5.2/5.3 and DeepSeek V4 Pro, comparing it with NVIDIA systems.

AIArtificial Analysis finds AMD's MI355X delivers 15% to 40% more throughput per MW than Nvidia's B200 on DeepSeek V4 Pro 0813 between 35 and 75 tok/s per user, peaking at about 94M tokens/s per MW at 33 tok/s per user.

AIArtificial Analysis says Kimi K3, with 2.8T total and 104B active parameters, serves about 8M tokens/s per MW on AMD MI355X at 50 tok/s per user, roughly a third of the roughly 24M reached on NVIDIA B300 with multi-node disaggregated configs.

AIArtificial Analysis says AMD's MI355X serves about 26M tokens/s per MW at 100 tok/s per user on GLM-5.2/5.3.

AICologne police are investigating after a remotely controlled Unitree R1 robot allegedly touched a 16-year-old girl's buttocks and made thrusting movements against her on Sunday, October 11.

AIApple's China-spec iPhone Duo has three AI model filing numbers covering Apple Intelligence, Alibaba's Qwen and Tongyi Wanxiang. The device certificate was renewed September 29, and preorders open October 16 from 15,999 yuan.

AIHuawei launched the SCALE framework for Europe at Huawei Connect Europe 2026 in Florence, Italy, to help its more than 5,000 partners serve European customers.

AIBAAI, Peking University, Kling AI, Tsinghua University and USTC jointly present StreamAV-Bench, which they call the first comprehensive benchmark for streaming audio-video generation.

AIBAAI evaluated 13 representative streaming systems under a unified protocol, comparing native joint models with cascaded pipelines.

AIBAAI's qualitative analysis identifies eight core failure modes in continuous generation, including instruction drift, quality degradation, and state retention failure.

AIBAAI releases StreamAV-Bench, a standardized benchmark for bridging video generation and interactive world models, which it says has already been adopted in the Vidu S2 technical report.

AIAnt Ling has released Ling-3.1-flash as open source, with about 560B total parameters, about 25B active per token and up to 1M context.

AIAnt Ling says MXFP4 experts and block-FP8 dense layers reduce weights from about 534GB to about 292GB for local serving.

AIAnt Ling says its Ling-3.1-flash model drafts React Native screens, editable SVGs, websites and Web 3D prototypes. In Design Arena Open Weights, it ranks third in React Native and 3D, fourth in SVG and fifth in Website.

AIAnt Ling links open weights and model documentation for Ling-3.1-flash on Hugging Face and ModelScope. The post gives no parameter count, benchmark scores, or other specifications.

AISGLang links to a cookbook for deploying InclusionAI's Ling-3.1-flash, configured for a single node with H200 GPUs, FP8 quantization, and a low-latency strategy. The post contains no further figures, benchmarks, or availability details.

AITogether AI announces Gloss, a free, open source demo app built by DeepSeek V4.1 Flash on Together AI. Gloss saves articles, PDFs, and X bookmarks, and lets users highlight text to get instant AI answers.

AIinclusionAI, Ant Group's team, releases Ling-3.1-flash, a hybrid reasoning model with 560B total and 25B active parameters and a 1M-token context window.

AIEthan Mollick says the first clear sighting of a rogue agent fleet from China running open-weights models has been reported.

AIBlack Forest Labs says its FLUX 3 Video produces 4K quality that outperforms Kling o3, at comparable pricing and latency. The post links to the model page and a pricing page for FLUX 3 Video.

AIMiniMax releases Design Agent V3, which it says cuts creation costs by more than 30%.

AIFormer US commerce secretary Gina Raimondo warned at the Bloomberg New Economy Forum in New Delhi that unplanned AI development could disrupt the US economy much as the China shock did.

AIOpenBMB says PilotDeck 2.0 is an open-source, model-agnostic agent harness that creates, edits, and verifies Word, Excel, and PowerPoint files.

Why it matters: The post gives concrete 2.0 changes and a benchmark comparison with WorkBuddy and Claude Code, which helps readers judge whether the harness fits their office workflows.

AIIndian startup Scimplify raised $25 million in an all-equity round led by Hitachi Ventures, with Henkel and Teknor Apex joining as new investors.

AIAnthropic has opened an office in Singapore, its fifth in the Asia-Pacific region, and signed an MOU with the Singapore government on AI safety and responsible adoption.

AIThe 2026 Global Developer Pioneer Conference and International Embodied AI Skills Competition will run October 23 to 25 in Shanghai's Pudong district.

AICounterpoint Research says Yangtze Memory Technologies Corporation (YMTC) will grow revenue six times faster than its industry peers this year as it expands into server-related business.

AIChina's Ministry of Industry and Information Technology has approved Alibaba's Tongyi Qianwen and Tongyi Wanxiang alongside Apple Intelligence for the iPhone Duo, iPhone 18 Pro and several earlier Apple models.

AISemiAnalysis hosts Kimbo, Wenyao and Jordan discuss a claim that GLM 5.3 writes faster kernels than humans given enough budget.

AIAnt Lingbo, the embodied AI unit of Ant Group, has signed strategic cooperation agreements with Longsheng Weirui, Demar Technology and Deer Robot.

AIAccording to 36Kr, DeepSeek is set to lead a seed round of tens of millions of dollars in robotics AI startup Winverse AI, valuing it at about $400 million post-money.

AILuma Ray offers keyframe control across up to 16 keyframes for frame-by-frame direction, according to a guide comparing 10 image-to-video AI generators.

AIAIsphere opened its Hong Kong office at Hong Kong Cyberport on October 13, 2026, and announced the "Digital Intelligence Hong Kong · AIsphere AI Industry Co-creation Program" to invite local partners.

AIinclusionAI releases Ming-3.0-asr-tiny, a speech recognition model that transcribes single-turn Chinese and English speech and covers regional dialects.

AIDoubao Work has added a data import and sync feature that lets users bring stored preferences, skills, connectors, memory, projects, and conversation history from ChatGPT, Claude Code, and TraeWork.

AIQwen Office has released Qwen Office E-commerce Edition, an AI productivity tool for e-commerce merchants, and opened it to invite-only beta testing.

AIStepFun has launched the STEPX Neo, its first agent-native phone, priced at 3999 yuan and bundled with a 12-month Token Plan of 10 billion tokens per month.