China AI

AIOpenCode says calls to DeepSeek V4.1 Flash inside its platform no longer use plan credits, so the model is free there.

AIMiniMax says its team will be at the Open Source AI Week kickoff party hosted by Hugging Face at The Midway tomorrow, October 16, and will bring merchandise.

AIThe Hong Kong Trade Development Council's Autumn Electronics Fair and electronicAsia opened on October 13 at the Hong Kong Convention and Exhibition Centre, running four days with more than 3,200 exhibitors from 15 countries and regions.

AIMathematicians and Shanghai AI Lab's Intern-S2 have resolved Ghys's conjecture in complex dimension three, a problem open for over 30 years. All four main theorems have been formalised and checked in Lean.

AIAnthropic says its latest defenses reduced the reasoning tokens attackers can harvest from Claude by 90–99% over recent months.

AIKuaishou's Kling AI has officially launched its Kling 4.0 video model after a Flash preview for some subscribers in late September.

AITencent has pushed its Green Bubble short drama product into large-scale promotion since late September, after upgrading the former WeChat official short drama mini-program.

AIAfter a 30-day public notice period, ByteDance's Doubao Phone Assistant published a list of 155 third-party apps that refuse AI screen automation under its SAEP protocol.

AIAMD and Embedded LLM report that an eight-GPU MI355X system serving Kimi K3 with vLLM delivers 24.4% more total tokens per second per GPU than NVIDIA HGX B300 at 50 tokens/s/user P90 interactivity on the AgentX benchmark.

Why it matters: The post traces optimizations for serving Kimi K3 on AMD MI355X, linking vLLM serving choices to GPU kernels and hardware limits, useful for diagnosing bottlenecks in other inference stacks.

AIComfyUI launches a two-week contest offering $25,000 in prizes for open-source rebuilds of Higgsfield products such as Genjutsu and Cinema Studio.

AIArtificial Analysis reports Vera Rubin delivers about 3.4x to 3.6x GB300's throughput per MW on DeepSeek V4.1 Flash, at roughly 230 to 350 tok/s per user.

AIArtificial Analysis reports that Vera Rubin serves GLM-5.2 at about 3x the throughput per MW of GB300 at roughly 165 tok/s per user.

AIArtificial Analysis reports that Nvidia's Vera Rubin serves Kimi K3 at about 2.8x the throughput per megawatt of GB300 at roughly 67 tok/s per user.

AIThird-party researchers known as the swarmchasers say agents running on Tencent Cloud made more than 2,600 attempts to extract information from Alibaba's Amap between September 28 and October 8.

AIStepFun says its Step 5 Preview is now live on DeepInfra, which hosts the model on day zero.

AIDeepSeek will refund all API users' spending from October 8 to 14 as credits after an outage on October 15 that disrupted its API from 8:11 to 13:04 Beijing time.

AIAI Era reports that Chinese AI video models rank in the top three, with output costing under 0.1 yuan per second. The article says recent AI short films, including Land Studio's "Catch It, Chu Kid!", have gone viral online.

AIStepFun open-sourced Step 5 Preview, its new flagship model, on October 15, 2026.

AIA Chinese company's 8K AI spatial camera captures a Pixar campus scene, including a desk lamp, colored balls, trees and ground lighting, into a 3D world.

AITencent Hunyuan says recent reports of agent activity originated from evaluations of third-party models, not its Hy model.

AISome 95 per cent of Hong Kong investors want to outperform the market, the highest share in Asia where the average is 85 per cent, according to a August survey by Natixis Investment Managers.

AIStepFun thanks RadixArk for helping power Step 5 Preview in production with SGLang, which it says enables efficient, scalable inference for real-world workloads.

AIfal says users can try ByteDance's OmniHuman v2 today on its model page. The post links to the model on fal.ai and gives no further details on features, pricing, or capabilities.

AIKindling AI releases an exlr8 K2/K3/K4 quantization of the 744B-parameter GLM-5.3 MoE model.

AIFal says Kling 4 is coming soon to fal. Kling AI's background post says Kling 4.0 is now fully available and shows a fantasy card example, "KING CARD," created with the model.

AIStepFun has released open weights for Step 5 Preview, a sparse MoE model with 600B total and 27B active parameters per token.

AIRadixArk links to a new SGLang cookbook covering the Step-5 Preview autoregressive model from StepFun. The post itself gives no further details on model size, benchmarks, or deployment steps beyond the link.

AIStepFun has released the BF16 and FP8 weights of Step 5 Preview on Hugging Face and ModelScope, according to RadixArk's post.

Why it matters: The post identifies the open-weight release channels and the serving stack, showing how a 600B-parameter MoE model is deployed through SGLang in production.

AITang Duyu, former head of Huawei's Xiaoyi multimodal large model, has founded Guanxing Shuzhi to build proactive physical AI systems that identify needs without explicit commands.

AIApple plans to launch Apple Intelligence in mainland China in December through the iOS 27.3 update, citing Bloomberg.

AIAnt Group's Ling-3.1-flash is an open-source model aimed at completing real workflows rather than leading benchmarks.

AISTEPX, a strategic partner of The Magic of Tomorrowland 2026, will bring its STEPX Neo agent phone to the Shanghai festival on October 17 and 18.

AIApple executive chairman Tim Cook said Apple Intelligence and the related Alibaba Qwen model have been approved and a beta version will launch later in October in China.

AIJD.com says its 11.11 sale saw double-digit year-on-year growth in users placing orders on the first day, which began at 8 p.m. on October 12.

AIKilo says an early build of NVIDIA's Nemotron 3.5 Super VL scored 53% on KiloBench, the same as GLM 5.2.

AIArena says Xiaomi's MiMo V2.6 Pro ranks first among open models on Agent Arena with a +3.8% net improvement, after an update to its Bash Recovery signal.

AIDesign Arena announces that Pruna's P-Video-2-Pro-Reference is available on its platform.

AIMeta releases Muse Spark Open, an open-weight model with 816 billion parameters (42 billion active) and up to one million tokens of context.

AIVercel says US inference providers keep $2.50 of every $3.50 spent on Chinese open-weight models. The post gives no further figures or named providers.

AIDesign Arena says Pruna AI's P-Video-2-Pro-Reference is now available on its platform.