China AI

AIStepFun's Step 5 Preview is now live on OpenRouter as the company's new flagship for agentic work.

AIShengshu Technology has opened a preview of its Vidu Q4 video generation model, which supports native 4K output and up to 15 reference images and three reference audio clips.

AITencent has added a new feature to its WorkBuddy product that lets users open files in different formats for preview and edit them with AI chat.

AIStepFun says Step 5 Preview is now available on OpenRouter, with a week of free access rolling out across OpenCode, Cline, Nous Research, Kilo Code, and other tools.

AIStepFun has released Step 5 Preview, its flagship model for agentic work, built on a sparse Mixture-of-Experts architecture with 27B active and 600B total parameters.

AITencent Cloud has open-sourced Octop, a self-hosted multi-agent AI assistant platform aimed at families and small teams, with multi-user accounts and data kept on the user's own machine.

AIAnalysts expect Alibaba Group Holding's cloud and AI unit to report a 50 per cent revenue surge for the September quarter, up from 45 per cent growth in the prior quarter.

AIQualcomm has denied being the net payer in its patent deal with Huawei and rejected claims that the deal relates to LogicFolding. The available statements do not establish any licensing of LogicFolding or related peripheral technologies.

AIvLLM v0.31.0 is released with 717 commits from 307 contributors, including 96 first-time contributors.

AIManus parent company Butterfly Effect has completed a new financing round of over $500 million, led by Boyu Capital and IDG Capital, and is rebuilding a team in Beijing to develop AI Agent products for the Chinese market.

AIAlibaba's official handbook on AI Native R&D identifies four open challenges: infrastructure engineering complexity, enterprise knowledge assets not yet agent-friendly, organizational design, and the pace of AI iteration.

AIGuizang (@op7418) says he posted news of Anthropic's price cut for Haiku 5.5 and was attacked by commenters who view the pricing as aimed at Chinese models.

AIA suspected Chinese-speaking attacker breached several South Korean financial institutions between late September and early October 2026, reportedly stealing over 25,000 records from Shinhan Bank alone.

Why it matters: The case shows how AI-driven penetration tools let one attacker breach several banks in a short window, a risk experts had warned about.

AIGeely Automobile Group launched its Geely Smart Charging technology on September 23, 2026, reaching a 2250 kW peak single-gun charging power and keeping maximum temperature at or below 65°C. The system, co-developed with StepFun and built on its PowerMind energy model, reportedly raises battery cycle life by more than 20% and targets county-level coverage by the end of 2027.

AIShanghai Biren Technology shares fell more than 12% in Hong Kong on Thursday after the company filed plans to raise HK$4 billion (US$510 million) through a share placement.

AILast week, several of South Korea's largest banks were hit by a cyberattack.

AIAnt Ling's Ling-3.1-flash is now available as a day-0 partner release on Vercel AI Gateway, free through October 13. The model is positioned as token-efficient, built for agentic and coding work, and suited to multi-step analysis.

AIHuawei released its first smartphone with its own LogicFolding chip, the Mate 90 series, on October 1, as China's smartphone and car markets slow.

AIAnt Ling announced a day-zero collaboration with AI/ML API, making Ling-3.1-flash available there for agentic and cowork scenarios.

AIAnt Ling announced a day-0 collaboration with Kilo Code for its launch, with Ling 3.1 Flash available for free in Kilo Code until October 13th. Novita Labs is serving the model.

AITencent's WorkBuddy agent can take a plain-language request through to a WeChat mini-program preview and a publish request, according to a hands-on product test reported on October 8.

AIA Galbot, Tsinghua University and collaborators paper won IROS 2026's Best Entertainment and Amusement Paper Award for LATENT, a humanoid tennis-return method trained on imperfect amateur motion-capture clips.

AIHuawei engineers presented XMFS, an experimental Linux kernel prototype filesystem, at the Linux Plumbers Conference in Prague on October 5.

AINathan Benaich's 9th annual State of AI Report covers agents, robotics, AI for science, inference economics, and government control over frontier AI access.

AIAnthropic has added simplified and traditional Chinese as on-screen language options in Claude's web interface in supported markets.

AIHu Han, former head of Tencent Hunyuan's visual large model algorithm center, is raising tens of millions of dollars for a multimodal startup at a target valuation of several hundred million dollars, with Yuanshi Capital as financial advisor.

AILing-3.1-Flash from Ant Ling is now available via Novita on OpenRouter, with Novita launching as a Day-0 partner.

AIOpenBMB releases MiniCPM-V-4.7-35B-A3B-DSpark, a 389,223,169-parameter DSpark draft model that pairs with MiniCPM-V-4.7-35B-A3B and proposes 15 tokens per forward pass.

AIOpenBMB has released MiniCPM-V-4.7-1B-DSpark, a 213,547,777-parameter draft checkpoint that pairs with the MiniCPM-V-4.7-1B target model and its tokenizer.

AIA South China Morning Post reporter spent a week using Huawei's Mate XT 2, a US$3,500 trifold phone with a 10.2-inch unfolded display.

AIByteDance ranked first in the debut SCMP Plus Going Global 100 ranking of mainland companies' overseas expansion.

AIAnthropic's Haiku 5.5 is priced below DeepSeek-V4.1 flash and Zhipu's GLM 5.3 flash. On Terminal Bench 4.0, the author's comparison shows it scoring one point above some rivals and level with GLM 5.3 flash, while ahead of the other two.

AIUBTECH Robotics and FAW-Volkswagen signed a strategic cooperation agreement to jointly develop and test embodied AI robot applications in logistics and build demonstration sites.

AIComfyUI announces a preview of Vidu Q4, inviting users to try it via a linked page. The post gives no further details on features, pricing, or availability terms.

AIComfyUI says Vidu Q4 Preview, the first preview of Vidu's new flagship video model, is now available through Partner Nodes.

AIInferact, working with DeepSeek, NVIDIA, and SemiAnalysis alongside the vLLM community, says joint work across models, custom kernels, and engine serving cuts time to first token (TTFT) by nearly 70% at ~100K throughput. vLLM is the open-source inference engine, and Inferact optimizes it for enterprise production deployments.

AIClaude's web and desktop apps now show a language selector with Chinese (Simplified) and Chinese (Traditional) options. The source describes the rollout as quiet and does not specify further details.

AIThe item is a Chinese media report headlined as Claude solving a probability theory problem, described as a step toward Fields Medal-level mathematics.

AITogether with DeepSeek's model and kernels and NVIDIA's collaboration, vLLM reports that time to first token (TTFT) drops nearly 70% at roughly 100K throughput.

AIvLLM has integrated DeepSeek's MegaAttention, which uses NVFP4 KV cache and is 45% smaller, alongside Mega-mHC, Mega-Gate, and DeepSelect kernels.