China AI news, from the source
Top stories
Updated Oct 13, 2026 Share this briefingToday's brief is being compiled. Verified Events keep landing in the live update ledger below and in the full Event History.
Browse the Event HistoryMore China AI news
Evidence-qualified product, pricing, policy, capital and deployment Events, ordered by when they happened. Research appears only after an impact review; unverified Signals stay separate.
20 of 39 stories · last 7 days
Newest first. Wider windows expand what is available; open Event History for the full period.
Today · Oct 13, 2026 11 stories
Reportedly: Deep Robotics is pursuing an IPO
Deep Robotics is pursuing an IPO, TMTPost reported. The report focuses on how the company plans to scale volume without relying on performance-oriented robots. No financial detail…
Why it matters If the IPO proceeds, Deep Robotics would gain access to public capital, potentially intensifying competition in China's robotics sector and affecting valuations of private robotics firms.
Reportedly: DeepSeek open-sources operator toolchain with native Huawei Ascend support
DeepSeek released open-source versions of its operator development tool TileLang and five core compute/communication libraries (DeepGEMM, DeepEP, FlashMLA, TileKernels, DeepSelect…
Why it matters Developers using Nvidia's CUDA ecosystem now have an open-source alternative that runs on Huawei Ascend NPUs, potentially lowering the cost and complexity of porting AI workloads to Chinese-made…
Reportedly: Tencent is developing a standalone 'Yuanbao Assistant' app for personal agent tasks
Tencent is developing a Personal Agent product called 'Yuanbao Assistant' with a standalone app, IT Home reported, citing people familiar with the matter. The app would run long-c…
Why it matters Tencent's reported move into a standalone personal agent app would give it a direct competitor to ByteDance's 'Xiaodou', potentially shifting consumer AI assistant competition toward long-term…
Reportedly: ModelBest releases MiniCPM-V 4.7, an on-device multimodal model with 13+ token/s decoding at 3GB memory
ModelBest released MiniCPM-V 4.7, an on-device multimodal model in 35B-A3B and 1B sizes. The 35B-A3B MoE version runs at 13+ token/s with 3GB memory on iPhone 17 Pro Max. It suppo…
Why it matters On-device model developers can now run a 35B-parameter MoE multimodal model on a flagship phone with 3GB memory, lowering the hardware barrier for local AI features.
01.AI publishes web-assets repository on GitHub
01.AI created a new public GitHub repository named web-assets. The source release was published on 2026-10-13. No further details about its contents or purpose are provided.
Why it matters Developers outside China gain a new public resource from 01.AI, but its utility is unknown until the repository contents are inspected.
Reportedly: DiffuSpace raises 500M yuan, Huawei and Horizon Robotics among investors
Shenzhen-based DiffuSpace is raising a 500 million yuan (approx. $70M) funding round, with Huawei and Horizon Robotics among the investors, TMTPost reported, citing people familia…
Why it matters Huawei and Horizon Robotics would gain exposure to an alternative language model architecture, potentially reducing dependence on autoregressive models for edge and automotive applications.
Reportedly: Former SenseTime VP's startup Feiyu Intelligence raises over ten million dollars within three months of founding for private enterprise agents
Feiyu Intelligence, founded by a former SenseTime vice president, raised over ten million dollars within three months of founding, TMTPost reported. The startup focuses on full-st…
Why it matters Enterprise buyers evaluating private AI deployments may see a new vendor option focused on full-stack private agents, potentially affecting procurement timelines and build-vs-buy decisions.
OpenBMB releases Trident repository for mixed-modality retrieval
OpenBMB (ModelBest) published the Trident repository on GitHub, introducing a method to mitigate modality preference in mixed-modality retrieval. The repository is under construct…
Why it matters Developers working on retrieval systems over mixed text-image corpora may face a new open-source option from OpenBMB once code and checkpoints are released, potentially reducing the need to build…
Reportedly: Infinigence AI raises RMB 2.3 billion Series C, setting China AI-native infrastructure record
Infinigence AI has completed a RMB 2.3 billion (approx. $320 million) Series C round, bringing cumulative funding to RMB 4.3 billion, IT Home reported. Both the single round and c…
Why it matters Infinigence AI's record funding round strengthens its position in China's AI infrastructure market, potentially increasing competition for compute orchestration and optimization services.
Reportedly: Zhipu AI Gray-Releases GLM-5.3-Highspeed in ZCode, Users Measure About 490 Tokens/s
Zhipu AI is quietly testing GLM-5.3-Highspeed with randomly selected ZCode subscribers, Pandaily reported. One user measured about 490 tokens/s; Zhipu has not published an officia…
Why it matters If confirmed, GLM-5.3-Highspeed's reported speed could pressure competitors to match or exceed it in high-throughput inference, potentially lowering costs for developers using ZCode.
Reportedly: Manus AI releases Manus 2.0 with Cascade framework, cloud computers, and Cue personal assistant
Manus AI released Manus 2.0, introducing the Cascade agent framework that cuts task token consumption by 23.2%, completion time by 28.2%, and running costs by 32%, GeekPark report…
Why it matters Manus AI's 2.0 release directly pressures competing agent platforms by bundling cloud computing, automation, and a personal assistant into one ecosystem, potentially reducing the need for users to…
Yesterday · Oct 12, 2026 9 stories
OpenBMB released Trident-Qwen3VL-2B, a multimodal embedding model for mixed-modality retrieval
OpenBMB released Trident-Qwen3VL-2B on Hugging Face, a multimodal embedding model fine-tuned from Qwen3-VL-2B-Instruct on Trident-Train. It maps text, images, and fused inputs int…
Why it matters Developers building retrieval systems for mixed text-image corpora can now use an open-weights model that ranks text, image, and fused documents in a single list, reducing the need to maintain…
Skywork open-sourced Matrix-Game-4.0 under MIT license
Skywork released Matrix-Game-4.0 on Hugging Face under the MIT license. The source release was published on 2026-10-12.
Why it matters Developers outside China gain immediate access to a new open-weights model under a permissive MIT license, reducing integration cost and legal friction for commercial use.
Reportedly: Huawei's HarmonyOS 7 Xiaoyi AI Quick Bar to roll out in batches by end of October
Huawei's HarmonyOS 7 Xiaoyi AI Quick Bar is expected to roll out in batches by the end of October, IT Home reported, citing Huawei customer service. The feature will keep Xiaoyi r…
Why it matters Huawei device users gain persistent, one-tap access to the Xiaoyi agent, which could increase usage of on-device AI features and shift interaction away from app-based assistants.
Reportedly: Chinese Tech Groups Shift AI Spending From Model Prices to Data Center Power and Depreciation
Chinese tech groups are pouring capital into data centers in power-rich regions such as Ulanqab as model APIs get cheaper, TMTPost reported. The next contest is who can keep serve…
Why it matters Chinese AI companies face rising depreciation and power costs for data centers, which will pressure margins unless server utilization stays high.
Reportedly: Moonshot AI seeks pre-IPO settlement with Recurrent.ai over shareholder dispute
Moonshot AI is negotiating with shareholders to settle a legal dispute before its IPO, IT Home reported. Recurrent.ai and some investors accuse CEO Yang Zhilin and CTO Zhang Yutao…
Why it matters Moonshot AI's IPO timeline and valuation could be affected by the shareholder dispute, as Recurrent.ai and its investors seek compensation equal to 10.5% of Moonshot's original shares.
Tencent Hunyuan open-sources OmniCapBench audio-visual captioning evaluation framework
Tencent Hunyuan released OmniCapBench, an open-source evaluation framework for fine-grained audio-visual captioning, on GitHub. The repository includes an annotation pipeline and …
Why it matters Researchers and developers evaluating audio-visual captioning models gain a reproducible, open-source benchmark that localizes errors to specific units, reducing reliance on opaque global LLM judges…
Reportedly: StepFun's Step 5 Preview Comes to Tencent WorkBuddy, Free to Use Exclusively Until October 25
StepFun's Step 5 Preview model is now available inside Tencent's WorkBuddy office agent, with an exclusive free-use window running from October 12 to October 25, Pandaily reported.
Why it matters Tencent WorkBuddy users gain free access to StepFun's Step 5 Preview model until October 25, potentially reducing their AI costs during the promotional period.
Ant Group releases PolyOCR-Venus OCR foundation models
Ant Group's GuangJian Team released PolyOCR-Venus, a family of 2B and 9B OCR foundation models built on Qwen3.5, with a technical report and OCRBench evaluation code. The release …
Why it matters Developers outside China cannot yet use PolyOCR-Venus because model weights are not released; they can only evaluate the OCRBench code and technical report.
Reportedly: StepFun Step 5 Preview model goes live on Tencent WorkBuddy with exclusive free access from October 12 to 25
StepFun announced that its Step 5 Preview model is now available on Tencent WorkBuddy, with exclusive free access from October 12 to 25, 2026, IT Home reported. The model uses a s…
Why it matters Tencent WorkBuddy users gain free access to a 600B-parameter model for two weeks, potentially lowering the cost of testing advanced AI capabilities in enterprise workflows.
Reportedly: Xiaomi launches Mijia Robot Vacuum-Mop 7 Max with 50°C hot-water mopping from RMB 4,999
Xiaomi announced the Mijia Robot Vacuum-Mop 7 Max, opening pre-sales on October 12, 2026, with a water-tank version at RMB 4,999 and a slim built-in water version at RMB 5,299, IT…
Why it matters Xiaomi's launch of a robot vacuum-mop with hot-water mopping at a mid-range price point pressures competitors in the smart home cleaning segment to match the feature or adjust pricing.
Oct 11, 2026 0 stories
Reportedly: Qwen-Image-2.1 Pro and Turbo APIs launch on Qianwen AI platform
Alibaba's Qianwen AI platform launched Qwen-Image-2.1 Pro and Turbo APIs on October 11, 2026, enabling developers to integrate image generation and editing without deploying model…
Why it matters Developers building design, e-commerce, or content creation tools can now integrate Qwen-Image-2.1 capabilities via API, reducing the need for in-house model deployment and inference management.
Reportedly: DeepSeek Harness v0.2.1-alpha.2 Pre-Release Adds Reasoning Translation and Git Worktrees
DeepSeek released Harness v0.2.1-alpha.2, an open-source agent runtime pre-release adding experimental reasoning translation and Git Worktrees plugins, a working_directory tool, S…
Why it matters Developers using DeepSeek Harness gain new experimental plugins and tools for reasoning translation, Git worktree management, and SSH access, potentially simplifying agent workflows.
Oct 10, 2026 0 stories
Reportedly: Tencent Yuanqi to shut down on November 9, 2026, making created agents unavailable
Tencent Yuanqi will stop service on November 9, 2026, making all created agents unavailable, IT Home reported. From October 29, 2026, new agent creation and API key distribution w…
Why it matters Developers and businesses using Tencent Yuanqi for agent deployment must migrate to alternative platforms before November 9, 2026, or risk losing access to their agents and data.
No additional verified events in this window. Try a wider period or open Event History.
View All eventsNew Chinese AI models
All models- XiaomiMiMo-V2.6-ProSep 21, 2026 · Open weights
- DeepSeekDeepSeek-V4.1-FlashSep 10, 2026 · Open weights · Global API
- Zhipu AIGLM-5.3-FlashAug 26, 2026 · Open weights · Global API
- AlibabaQwen3.8-Flash-NextAug 26, 2026 · Open weights · Global API
- Zhipu AIGLM-5.3Aug 18, 2026 · Open weights · Global API
- AlibabaQwen3.8-27BAug 14, 2026 · Open weights · Global API
Explore China AI
- China AI modelsAccess, open weights, and licenses across Chinese labs.
- Chinese AI companies73 frontier labs, big tech, robotics and AI chip makers.
- China AI policyRules, algorithm filings, labeling standards and export controls.
- Attention IndexEnglish and Chinese source attention.
- China AI ExplainersEssential primers on terminology, frontier startups, and compliance.
- China AI LinesFive strategic themes: open weights, price wars, robotics, chips and policy.
Your reading list
Save stories. Free, across devices.