Event History
Chinese AI events, with original sources.
Filter by domain
Start with all, official, or research. Expand to filter by domain.
Reportedly: Underdog AI releases Saluki 27B: a 2-bit quantized Qwen3.8-27B at 7.89 GB
Underdog AI released Saluki 27B, a 2-bit quantized version of Qwen3.8-27B with a file size of 7.89 GB, IT Home reported on October 10, 2026. The model is designed for agent applications and excels at tool use, outperforming the full-size Qwen3.8-27B in tool selection and parallel tool calling. It uses standard llama.cpp format and loads without custom compilation.
NEW · Updated Oct 17, 2026 · TencentReportedly: Tencent plans up to $5 billion offshore bond, possibly to fund AI push
Tencent is considering issuing up to $5 billion in offshore bonds, possibly in US dollars and offshore yuan, as soon as this month, Leiphone reported citing people familiar with the matter. The proceeds may be used to increase investment in AI.
NEW · Updated Oct 17, 2026 · KuaishouReportedly: Kuaishou's Kling AI picks banks for Hong Kong IPO of at least $1 billion
Kuaishou's Kling AI has selected underwriters for a Hong Kong IPO aiming to raise at least $1 billion, IT Home reported, citing people familiar with the matter. The company is working with CICC, Goldman Sachs and UBS, and targets a listing as early as next year.
NEW · Updated Oct 17, 2026 · DeepSeekReportedly: DeepSeek web and API services suffer partial outage with 'server temporarily unavailable' errors
DeepSeek's web and API services experienced a partial outage on October 15, 2026, with users seeing 'server temporarily unavailable' errors. The official status page confirmed the disruption, which had lasted over 50 minutes at the time of reporting, IT Home reported.
NEW · Updated Oct 17, 2026 · Purple Mountain Laboratories and China MobileReportedly: Purple Mountain Laboratories and China Mobile Launch AXP Agent Interconnection Platform
Purple Mountain Laboratories and China Mobile launched the AXP Agent Interconnection Platform, built on deterministic networking, designed to register one million AI agents and run 100,000 concurrently. Cross-region agent addressing was demonstrated over 4,000 kilometers on the CENI testbed, Pandaily reported.
NEW · Updated Oct 17, 2026 · Shanghai AI LaboratoryReportedly: Shanghai AI Laboratory Releases Atria Dawn Full Version With a 512K Context Window
Shanghai AI Laboratory released the full version of its Atria Dawn agent model, built on a 744B MoE GLM-5.2 base, with a 512K context window. The release improves scores on JobBench, WideSearch, and PaperBench over the September preview. Full weights are still marked 'Coming soon', Pandaily reported.
NEW · Updated Oct 17, 2026 · TencentReportedly: 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 will stop. Users can export data until January 9, 2027.
NEW · Updated Oct 17, 2026 · DeepSeekDeepSeek released DeepSeek-V4.1-Flash-1017 on Hugging Face
DeepSeek released DeepSeek-V4.1-Flash-1017 on Hugging Face, a minor update to DeepSeek-V4.1-Flash with improved reasoning, coding, agentic, and visual benchmarks. The model is licensed under MIT and uses the transformers library.
NEW · Updated Oct 17, 2026 · TencentReportedly: 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-cycle tasks and personal long-term memory independently of the main Yuanbao app. Tencent Yuanbao, based on the Hunyuan large model, launched on May 30, 2024.
NEW · Updated Oct 17, 2026 · Unitree RoboticsReportedly: Unitree Releases UniUni Swarm Control Software, Free for Its Users, Starting With H2 and G1
Unitree Robotics released UniUni, a free desktop swarm control app for Windows and macOS, initially supporting its H2 and G1 humanoids. Users can draw paths, arrange timelines, sync motion to music, monitor fleets, and trigger a one-tap emergency stop, Pandaily reported.
NEW · Updated Oct 17, 2026 · TencentTencent Hunyuan released Hy-Memory-30B-A3B, a compact long-term-memory writer model
Tencent Hunyuan released Hy-Memory-30B-A3B, a ~3B active parameter long-term-memory writer model, on Hugging Face under Apache-2.0. It replaces general-purpose LLMs in Mem0, MemOS, and Hy-Memory frameworks. The model is trained via memory on-policy distillation and attributed memory reinforcement learning.
NEW · Updated Oct 17, 2026 · DeepSeekReportedly: 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, SDK session-directory APIs, and an SSH helper, Pandaily reported.
NEW · Updated Oct 17, 2026 · AlibabaReportedly: Alibaba Qwen open-sources BlenderWorld, a 223-task benchmark for 3D creation agents in Blender
Alibaba Qwen released BlenderWorld on October 14, an MIT-licensed benchmark with 223 tasks and 2,651 criteria for evaluating visual agents on real 3D creation in Blender. Code and data are on GitHub, Hugging Face, and ModelScope, IT Home reported.
NEW · Updated Oct 16, 2026 · TencentTencent open-sourced DSpine, a parallel drafter for Qwen3-8B speculative decoding
Tencent released Qwen3-8B-DSpine-b16 on Hugging Face under Apache 2.0. The model is a parallel drafter for speculative decoding, achieving an average acceptance length of 7.83 at temperature 0 on Qwen3-8B. Inference code was released in October 2026.
NEW · Updated Oct 16, 2026 · BaiduReportedly: Apple to launch Apple Intelligence in mainland China via iOS 27.3 in December with Baidu and Alibaba as suppliers
Apple plans to launch Apple Intelligence in mainland China via iOS 27.3 in December, IT Home reported, citing Bloomberg. The rollout covers iPhone, iPad, and Mac with AI features first released in the US two years ago, excluding the new Siri AI assistant. Baidu will provide search and Siri-related functions, while Alibaba is also named as a supplier.
NEW · Updated Oct 16, 2026 · TencentReportedly: Tencent Cloud open-sources TeamAI, a Git-based tool for sharing skills across agents
Tencent Cloud open-sourced TeamAI, an internal Git-based team AI collaboration tool that stores skills, work rules, tool configs, and project knowledge in a Git repo for review, update, and versioning, then syncs them to team members' agents. It supports 16 agents including WorkBuddy, CodeBuddy, Claude Code, Codex, and Cursor. TeamAI also added unified model configuration distribution for admins to enforce team security requirements, IT Home reported.
NEW · Updated Oct 16, 2026 · Ministry of Commerce of the People's Republic of ChinaReportedly: China proposes adding AI, biotech, and solar technologies to export control list
China's Ministry of Commerce and Ministry of Science and Technology issued Announcement No. 48 of 2026, proposing to add 40 key technologies in AI, biomedicine, and photovoltaics to the export control list, IT Home reported. The proposal includes 14 technologies proposed for export prohibition and 26 for export restriction, with public comments being solicited.
NEW · Updated Oct 16, 2026 · 4ParadigmReportedly: 4Paradigm and RiseUnion Open-Source ModelSphere for Heterogeneous LLM Serving
4Paradigm and RiseUnion open-sourced ModelSphere, an LLM deployment tool unifying cluster serving, KV caching, and AutoTune under Apache 2.0. It claims support for Ascend, Iluvatar, Moore Threads, and Hygon DCU, Pandaily reported.
NEW · Updated Oct 16, 2026 · AgiBotReportedly: AgiBot starts new funding round at about RMB 50 billion valuation
AgiBot is raising RMB 1–1.5 billion in a new round nearing completion at a valuation of about RMB 50 billion, IT Home reported, citing Caijing. The round means AgiBot cannot complete an IPO this year under HKEX rules.
NEW · Updated Oct 16, 2026 · Ant GroupinclusionAI released LemonDecider, a 4B decision model with explicit abstention
inclusionAI released LemonDecider, a 4B-parameter decision model derived from Qwen3.5-4B, on Hugging Face under Apache-2.0. It uses a 255-way single-token readout for choice, noul, and score questions with explicit abstention. The model is available for download with custom code and a standalone CLI.
NEW · Updated Oct 16, 2026 · DeepSeekReportedly: OpenCode makes DeepSeek V4.1 Flash free with 200K context
OpenCode announced DeepSeek V4.1 Flash is now free, with the free tier supporting a 200K context window, down from the 1M version, IT Home reported. The move follows earlier quota increases under 'Operation Cheepseek'.
NEW · Updated Oct 16, 2026 · StepFunReportedly: 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.
NEW · Updated Oct 16, 2026 · XiaomiReportedly: Xiaomi open-sources MiDashengLM-Spatial, an end-to-end spatial audio understanding large model
Xiaomi released MiDashengLM-Spatial, an open-source end-to-end spatial audio understanding large model, with weights and inference code under Apache 2.0. It builds on MiDashengLM-7B and adds spatial perception and reasoning. The model will be applied in Xiaomi smart home and car cockpit scenarios, IT Home reported.
NEW · Updated Oct 16, 2026 · KuaishouReportedly: Kuaishou Kling 4.0 video model fully launched, supports up to 30-second generations
Kuaishou's Kling AI announced the full launch of Kling 4.0 on klingai.com, with single video generations up to 30 seconds. 4K and 1080p 10-bit HDR output is coming soon, and API access for enterprises and developers will open in the near future, IT Home reported.
NEW · Updated Oct 16, 2026 · MetaReportedly: Meta releases Muse Spark Open, an 816B-parameter open-weights multimodal model
Meta released Muse Spark Open, an open-weights model with 816B total parameters, 42B active per token, and 1M-token context. It supports multimodal input and targets coding and agent tasks. The model is available via API platforms and downloadable from Hugging Face, Zhidongxi reported.
NEW · Updated Oct 15, 2026 · DeepSeekReportedly: 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) with native support for Huawei Ascend 950 NPU, Leiphone reported on 2026-09-30.
NEW · Updated Oct 15, 2026 · TencentTencent 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 scoring toolkit. The source release was published on 2026-10-12.
NEW · Updated Oct 15, 2026 · ByteDanceReportedly: ByteDance's Doubao Phone Assistant blocked from GUI automation by 151 apps including WeChat, Taobao, and Meituan
ByteDance's Doubao Phone Assistant will not perform GUI automation on 151 apps that declined under the Screen Automation Operation Statement Protocol, including WeChat, Taobao, and Meituan, Leiphone reported. The 30-day public comment period has ended, and developers can still opt out. The restriction applies only to automated operations, not app installation or manual use.
NEW · Updated Oct 15, 2026 · Manus AIReportedly: 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 reported. The update adds cloud computers, automation triggers, Manus Studio, and the new Cue personal assistant app for mobile and desktop.
NEW · Updated Oct 15, 2026 · AppleReportedly: Apple Intelligence to integrate Alibaba Qwen large model, test version in China by late October
Apple's Tim Cook said Apple Intelligence and Alibaba's Qwen large model have been approved and a test version will be released in China in late October, Leiphone reported. The test version includes real-time translation, writing tools, and photo cleanup, but not Siri AI.
NEW · Updated Oct 15, 2026 · Shanghai AI LaboratoryReportedly: Shanghai AI Lab Open-Sources Intern-Decision, Small Models That Output Decisions, Not Text
Shanghai AI Lab open-sourced Intern-Decision, a family of small models in 0.8B, 2B and 4B sizes that output decisions, scores and yes/no answers with probabilities. MetaX shipped Day-0 support, Pandaily reported.
NEW · Updated Oct 15, 2026 · HuaweiReportedly: 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 familiar with the matter. The round would set a global record for diffusion language model funding.
NEW · Updated Oct 15, 2026 · Kunlun TechReportedly: Kunlun Tech releases Matrix-Game 4.0, the first video large model game running on a single consumer GPU, and open-sources it
Kunlun Tech released Matrix-Game 4.0, the first video large model game running on a single consumer GPU, achieving 480p at 60 FPS with under 150 ms latency on an NVIDIA RTX 5090. The project is open-sourced on GitHub and Hugging Face, IT Home reported.
NEW · Updated Oct 15, 2026 · AgiBotAgiBot open-sources VINE, a value-gradient fine-tuning method for flow-matching policies
AgiBot released the VINE codebase on GitHub under Apache License 2.0. VINE enables stable end-to-end value-gradient fine-tuning of flow-matching policies by reconstructing interpolation states at each denoising step. The repository targets Linux, Python 3.12, CUDA 12, JAX 0.5.3, and OGBench 1.2.1.
NEW · Updated Oct 15, 2026 · ByteDanceReportedly: ByteDance's Doubao phone assistant faces rejection from 155 apps over screen automation
ByteDance's Doubao phone assistant saw 155 third-party apps reject its AI screen automation after the SAEP protocol's 30-day public comment period ended on October 14, GeekPark reported. Tencent, Alibaba, Meituan, Kuaishou, Kimi, and DeepSeek products are among those refusing automation. The protocol lets developers decide whether to allow Doubao's screen recognition and simulated clicks.
NEW · Updated Oct 14, 2026 · ByteDanceReportedly: ByteDance's Doubao Assistant opens third-party app operation permissions
ByteDance's Doubao Assistant has opened permissions for third-party app operations, IT Home reported on October 15, 2026. The change was announced in a daily tech briefing. No further details on scope or availability were provided.
NEW · Updated Oct 14, 2026 · AlibabaQwenLM releases BlenderWorld benchmark for 3D creation agents
QwenLM released BlenderWorld, an open-source benchmark with 223 tasks across 9 categories for evaluating visual agents on 3D creation in Blender. The dataset includes 2,651 binary criteria and supports text, image, and video references. The source release was published on 2026-10-14.
NEW · Updated Oct 14, 2026 · AlibabaReportedly: 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 models. Qwen-Image-2.1-Turbo model weights were also open-sourced, IT Home reported.
NEW · Updated Oct 14, 2026 · Yangtze Memory Technologies Corporation (YMTC)Reportedly: YMTC to grow revenue six times faster than peers as it expands server business ahead of IPO
Yangtze Memory Technologies Corporation (YMTC) is set to grow revenue six times faster than industry peers this year, Counterpoint Research reported. Server-related revenue was effectively zero in 2024 but could reach nearly 39% of revenue by 2028. The company is expanding into server business ahead of a planned IPO.
NEW · Updated Oct 14, 2026 · BaiduPaddlePaddle releases VHA open-source repository
PaddlePaddle (Baidu) published a new open-source repository named VHA on GitHub. The source release was published on 2026-10-14. No further details about its contents or purpose are provided in the evidence.
NEW · Updated Oct 14, 2026 · ByteDanceReportedly: Doubao Work adds an infinite creation canvas with Seedream 5.0 Flash and Doubao 2.1 Lite
ByteDance's Doubao Work has added an infinite creation canvas powered by the new Seedream 5.0 Flash image model and Doubao 2.1 Lite, a lighter model for everyday office tasks, Pandaily reported.
NEW · Updated Oct 14, 2026 · TencentReportedly: Tencent WorkBuddy launches standalone file browser for local AI file processing
Tencent announced its AI office product WorkBuddy now has a standalone file browser, IT Home reported. Users can right-click a file in Explorer/Finder, open it with WorkBuddy, and use AI to analyze, modify, or continue processing. The feature supports Word, Excel, PPT, PDF, Markdown, and HTML files, with multi-tab switching and per-file AI conversations.
NEW · Updated Oct 14, 2026 · ModelBestReportedly: 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 supports iOS, Android, HarmonyOS, and chips from Intel, AMD, Arm, Rockchip, and Houmo, IT Home reported.
NEW · Updated Oct 14, 2026 · SenseTimeReportedly: 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-stack private enterprise agents. The report cites people familiar with the matter.
NEW · Updated Oct 13, 2026 · Deep RoboticsReportedly: 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 details or timeline were disclosed.
NEW · Updated Oct 13, 2026 · 01.AI01.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.
NEW · Updated Oct 13, 2026 · ModelBestOpenBMB 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 construction; code, data, and model checkpoints are not yet available.
NEW · Updated Oct 13, 2026 · Infinigence AIReportedly: 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 cumulative total are records for China's AI-native infrastructure sector.
NEW · Updated Oct 13, 2026 · Zhipu AIReportedly: 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 official speed figure.
NEW · Updated Oct 12, 2026 · ModelBestOpenBMB 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 into one shared space for mixed-modality retrieval. The model is available under the MIT License.
NEW · Updated Oct 12, 2026 · Kunlun TechSkywork 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.
NEW · Updated Oct 12, 2026 · HuaweiReportedly: 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 resident in the Dock bar at the bottom of the phone desktop.
NEW · Updated Oct 12, 2026 · TMTPostReportedly: 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 servers utilized and recover multi-year infrastructure bets.
NEW · Updated Oct 12, 2026 · Moonshot AIReportedly: 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 of misrepresentation and seek compensation equal to 10.5% of Moonshot's original shares.
NEW · Updated Oct 12, 2026 · Ant GroupAnt 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 includes benchmark results on OCRBench v2.1, CC-OCR, and OmniDocBench v1.6, but model weights and training code are not included.
NEW · Updated Oct 12, 2026 · TencentReportedly: 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 sparse MoE architecture with 600B total parameters, 27B active parameters, and a 1 million token context window.
NEW · Updated Oct 12, 2026 · XiaomiReportedly: 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 Home reported.
NEW · Updated Oct 11, 2026 · Shanghai AI LaboratoryOpenGVLab releases RSI-Nav, an open-source recursive self-improvement framework for vision-language navigation
OpenGVLab released RSI-Nav, an open-source recursive self-improvement framework for continuous vision-language navigation built on Qwen3.5-9B. The repository provides training, trajectory conversion, self-improvement, and R2R-CE evaluation code. Model weights and datasets are prepared separately.
NEW · Updated Oct 10, 2026 · AlibabaReportedly: Alibaba's Qoder Adds Fast Mode With 3-Second First Responses and Lower Credit Use
Alibaba's Qoder coding agent launched a Fast tier that answers in about three seconds at 0.8x credits, replacing the Performance tier on its international edition. The source announcement was published on 2026-10-10, Pandaily reported.
NEW · Updated Oct 10, 2026 · ShengShu AIReportedly: ShengShu AI releases Vidu Q4 Preview video generation model with 2K and 4K output
ShengShu AI released Vidu Q4 Preview, a video generation model supporting up to 3 reference audio clips, 15 reference images, and 2K/4K output. It offers image-to-video and reference-to-video generation, with a launch price of 0.09 yuan per second, IT Home reported.
NEW · Updated Oct 10, 2026 · ByteDanceReportedly: ByteDance's Doubao pricing strategy influences US AI price competition
ByteDance's Doubao large model pricing strategy is reportedly influencing US AI price competition, with Claude cutting prices to 10 cents and GPT offering free tiers, TMTPost reported.
NEW · Updated Oct 9, 2026 · XiaomiReportedly: Xiaomi retires MiMo-V2.5-Pro and MiMo-V2.5, auto-switching to V2.6 models on October 14
Xiaomi announced the deprecation of MiMo-V2.5-Pro and MiMo-V2.5, with automatic switching to MiMo-V2.6-Pro and MiMo-V2.6-Flash starting October 14, 2026, 18:00 UTC+8. Old model names will fail after October 21, 2026, 10:00 UTC+8, IT Home reported.
NEW · Updated Oct 9, 2026 · HuaweiReportedly: Huawei launches Smart Door Lock 2 Pro Enjoy Edition with AI 3D face recognition 3.0
Huawei launched the Smart Door Lock 2 Pro Enjoy Edition on October 9, 2026, with AI 3D face recognition 3.0 and a subsidized price of 2,124.15 yuan (approx. $296), IT Home reported. The lock unlocks in 0.8 seconds, 50% faster than the Huawei Smart Door Lock Pro, and captures 307,200 facial depth points using 3D ToF technology. Availability outside China is not stated in the evidence.
NEW · Updated Oct 9, 2026 · Amazon Web ServicesReportedly: Amazon Bedrock AgentCore launches in China regions
Amazon Web Services announced Amazon Bedrock AgentCore is generally available in the China (Beijing) and China (Ningxia) regions as of September 29, Zhidongxi reported. The platform offers six core capabilities: runtime, gateway, identity management, browser tools, code sandbox, and observability. Users in China can access it via AWS.
NEW · Updated Oct 9, 2026 · AlibabaQwen open-sourced Qwen-Image-2.1-Turbo, an 8-step accelerated image model
Qwen released Qwen-Image-2.1-Turbo, an accelerated checkpoint of Qwen-Image-2.1 for text-to-image generation and image editing with 8 denoising steps. It uses the same 7B visual generation architecture and loads directly with QwenImage21Pipeline in Diffusers. The checkpoint includes its recommended sampling schedule and is available on Hugging Face.
NEW · Updated Oct 8, 2026 · AlibabaReportedly: Alibaba’s AI cloud revenue surged over 50% in September quarter
Alibaba’s AI Cloud and Compute Services unit revenue grew over 50% year-on-year in the September quarter, up from 45% in the prior quarter, SCMP Tech reported citing analysts. The growth reflects returns on aggressive AI investments.
NEW · Updated Oct 8, 2026 · EcosiaReportedly: Ecosia drops Mistral for Chinese open models Qwen, GLM and Kimi
German search engine Ecosia has switched from Mistral to Chinese open models Qwen, GLM and Kimi via partner Melious, cutting costs by half and gaining performance, IT Home reported. The move follows Ecosia's May switch from OpenAI to Mistral, which suffered server overloads and quality issues.
NEW · Updated Oct 8, 2026 · TencentReportedly: Manus AI raises over $500M, Tencent among backers
Manus AI's parent Butterfly Effect has completed a new funding round of over $500 million (approx. $3.352 billion RMB), Zhidongxi reported. The round was led by Boyu Capital and IDG Capital, with existing investors Tencent, Sequoia China, and ZhenFund participating.
NEW · Updated Oct 8, 2026 · ModelBestOpenBMB released MiniCPM-V-4.7-1B-DSpark, a speculative decoding draft model for MiniCPM-V-4.7-1B
OpenBMB released MiniCPM-V-4.7-1B-DSpark on Hugging Face, a 5-layer draft model with 213,547,777 parameters trained for exact pairing with MiniCPM-V-4.7-1B and its tokenizer. It is licensed Apache-2.0 and supports vLLM. The source release was published on 2026-10-08.
NEW · Updated Oct 7, 2026 · Moonshot AIReportedly: Moonshot AI completes final pre-IPO funding round at about $50 billion valuation, plans Hong Kong IPO
Moonshot AI has completed its final private funding round before an IPO at a valuation of about $50 billion, IT Home reported, citing people familiar with the matter. The company plans to list in Hong Kong in the first quarter of next year, with a fundraising cap of $5 billion.
NEW · Updated Oct 6, 2026 · Shanghai AI LaboratoryShanghai AI Laboratory open-sources SpatialASM framework for spatial reasoning in LVLMs
Shanghai AI Laboratory's InternLM team released the SpatialASM repository on GitHub, introducing a framework that internalizes Active Spatial Mentalization into large vision-language models. The repository currently contains only a README; paper, models, data, and code are marked as coming soon.
NEW · Updated Oct 6, 2026 · DeepSeekReportedly: DeepSeek finalizes at least RMB 80B round with Tencent and CATL
DeepSeek is close to finalizing a new funding round of at least RMB 80 billion, with Tencent and CATL among the largest committed investors, Zhidongxi reported citing Bloomberg. The final total could approach RMB 100 billion, paving the way for a planned IPO in early 2027.
NEW · Updated Oct 4, 2026 · DeepSeekReportedly: DeepSeek Harness v0.2.1-alpha.1 Adds Experimental Claude Code Mods Compatibility Layer
DeepSeek Harness v0.2.1-alpha.1 adds an experimental bridge that runs Claude Code Mods as Harness plugins, plus a new entry for agent-built plugins and a developer toolkit, Pandaily reported.
NEW · Updated Oct 4, 2026 · HuaweiReportedly: Huawei's openJiuwen Open-Sources X-Router to Pick the Right Model for Each Agent Request
Huawei's openJiuwen open-sourced X-Router, a self-evolving model router that sends easy agent requests to light models and hard ones to stronger models, learning from feedback. Tests show near-baseline scores at roughly half the cost, Pandaily reported.
Updated Oct 3, 2026 · HuaweiReportedly: Huawei discontinues Xiaoyi Help agent support for new devices in HarmonyOS 7.0
Huawei's Xiaoyi Help agent will no longer support devices added after HarmonyOS 7.0, IT Home reported on October 3, 2026. Devices supported under HarmonyOS 6.0 can continue using the agent, and personal skills and scheduled tasks remain usable after upgrading to HarmonyOS 7.0.
Updated Oct 2, 2026 · HuaweiReportedly: Arcfox Alpha T5 OTA adds Apple CarPlay, Huawei HiCar and end-to-end 4.0 driver-assist model
Arcfox pushed an OTA update for the Alpha T5 adding Apple CarPlay, Huawei HiCar and an end-to-end 4.0 driver-assist model for Yuanjing trim vehicles, IT Home reported on Oct 2, 2026. The update includes five new features and one optimization. CarPlay and HiCar are exclusive to Yuanjing trim; other trims get CarLink.
Updated Oct 2, 2026 · GalbotReportedly: Galbot Prices ET1 Bipedal Humanoid From 79,000 Yuan, With Watch-and-Learn Skills on Upper Tiers
Galbot's ET1 bipedal humanoid is priced from 79,000 yuan across three versions (79,000, 139,000 and 179,000 yuan), aimed at performances, theme parks and storefront interaction, Pandaily reported.
Updated Oct 2, 2026 · AlibabaReportedly: Cloudflare launches open-source multimodal decision model Clef based on Qwen
Cloudflare launched Clef and Clef-flash, open-source multimodal decision models based on Qwen3.8-27B and Qwen3.5-9B, IT Home reported on October 2, 2026. Clef leads the Jev Decision Index 0.2.1 benchmark and extends context window to 64k.
Updated Oct 2, 2026 · MiniMaxMiniMax open-sources MiniMax-Code-MiniApps repository for community-built plugins
MiniMax released the MiniMax-Code-MiniApps repository on GitHub, an open-source collection of community-contributed MiniApp plugins for the MiniMax Code desktop app. The repository includes plugins such as Token Usage Board, Model Manager, and Self-drive Route Planner, and is open for pull requests.
Updated Oct 1, 2026 · AISphereReportedly: AISphere Launches PixVerse R2, a Real-Time World Model That Remembers What Happens in a Session
AISphere launched PixVerse R2, a real-time world model that generates explorable audiovisual worlds from text, image, audio, and keyboard input while maintaining state across a session. It is now fully available on the web. The launch was reported by Pandaily and Leiphone.
Updated Oct 1, 2026 · AlibabaReportedly: Bilibili Open-Sources Index-Translate, a Qwen3.5-Based Translation Model Family for 150 Languages
Bilibili's Index LLM team released Index-Translate under Apache-2.0: 2B, 9B and 35B-A3B preview text models for 150 languages, plus dubbing, syllable-control and long-document variants, Pandaily reported.
Updated Oct 1, 2026 · Horizon RoboticsReportedly: Horizon Robotics Releases HSD V2.1 With End-to-End Reversing, iCAR V27 First in Line for OTA
Horizon Robotics released HSD V2.1, adding end-to-end reversing in all scenarios, with 43% more miles per intervention and 255% better interactive lane changes. iCAR V27 OTA starts October 8, Pandaily reported.
Updated Oct 1, 2026 · HuaweiReportedly: Huawei Launches Grid-Interactive AIDC 1.0, Linking AI Data-Center Power, Cooling and Ascend Servers
Huawei Digital Power launched Grid-Interactive AIDC 1.0, pairing grid-forming storage, new UPS lines and an Ascend-aware liquid-cooling controller. A SenseTime site received 18 MW of power in 45 days, Pandaily reported.
Updated Oct 1, 2026 · Manus AIManus AI Launches Video Editor in Manus 2.0
Manus AI introduced a Video Editor in Manus 2.0, enabling anyone to make a video worth posting. The announcement was published on the Manus Blog.
Reportedly: Ant Group's Ling-3.1-flash launches with 560B total parameters and 1M context
Ant Group's Ling-3.1-flash model launched with about 560B total parameters and 25B active parameters per token, IT Home reported. The model has a 1M context window, but free trial service is limited to 256K for two weeks. After the free period, the team plans to open source the model and enable 1M context.
Updated Sep 30, 2026 · TencentTencent Hunyuan releases ExplorationBench, an open-source benchmark for AI exploration in verifiable alien worlds
Tencent Hunyuan released ExplorationBench, an open-source benchmark with two verifiable alien worlds, 55 discovery targets, 140 held-out tasks, and results for 10 frontier AI systems. The repository is available on GitHub, with a paper on arXiv and an official website.
Updated Sep 30, 2026 · AlibabaReportedly: Peking University, Tsinghua and Alibaba Open-Source SparkDiffusion for Up to 265x Faster Wan Video Generation
Peking University, Tsinghua and Alibaba open-sourced SparkDiffusion, which cuts a 720P Wan2.1 14B video from 4,769 to 18 seconds on one RTX 5090, a 265x speedup. Code and weights are open, Pandaily reported.
Updated Sep 30, 2026 · DeepSeekReportedly: DeepSeek Open-Sources Ascend Versions of TileLang, DeepGEMM and DeepEP as Huawei Details SuperPoD Flex
DeepSeek released Ascend versions of TileLang, DeepGEMM, DeepEP, TileKernels, FlashMLA and DeepSelect, while Huawei detailed a jointly defined 128-card SuperPoD Flex design. The source announcement was published on 2026-09-30, Pandaily reported.
Updated Sep 30, 2026 · ByteDanceReportedly: ByteDance names Doubao personal assistant 'Xiaodou' and plans standalone app
ByteDance's Doubao personal assistant product is reportedly called 'Xiaodou' and will get a standalone app, IT Home reported on September 30, 2026, citing media outlet Dujia. The name was decided in summer 2026, and the product is still in internal testing.
Updated Sep 30, 2026 · DeepSeekDeepSeek open-sources dsh-libreoffice-kit 0.1.0 for font-friendly Office conversion and rendering in Node.js
DeepSeek released dsh-libreoffice-kit 0.1.0, an open-source Node.js library for Office format conversion, spreadsheet recalculation, and PNG rendering. It requires Node.js 22.19.0 or newer and supports binary .doc, .xls, and .ppt files. The package installs the dsoffice command and uses native or WASM engines.
Updated Sep 30, 2026 · ByteDanceReportedly: ByteDance's Doubao adds one-stop travel booking and ride-hailing
ByteDance's Doubao assistant now supports flight and train booking, ride-hailing and navigation ahead of China's National Day holiday, Leiphone reported. The features are available in the Doubao app through partnerships with Flight Manager, High-speed Rail Manager, Cao Cao Mobility and Baidu Maps.
Updated Sep 30, 2026 · DeepSeekReportedly: DeepSeek open-sources Ascend base components with Huawei
DeepSeek has open-sourced Ascend base components to build an efficient AI chip software ecosystem with Huawei, QbitAI reported. The report cites no named sources. Availability of the components outside China is not stated.
Updated Sep 30, 2026 · DeepSeekReportedly: DeepSeek open-sources Huawei Ascend infrastructure components matching its NVIDIA stack
DeepSeek open-sourced infrastructure components for Huawei Ascend, including TileLang compiler, compute libraries, and distributed communication libraries, IT Home reported. The components correspond one-to-one with its NVIDIA platform releases. TileLang Ascend version wraps Ascend C instructions without hardware performance loss.
Updated Sep 30, 2026 · DeepSeekReportedly: DeepSeek Harness v0.2 Preview Adds Desktop Installers, Plugin Manager and Scheduled Automations
DeepSeek released Harness v0.2 preview with official macOS and Windows installers, a plugin manager that installs npm packages without a command line, file previews and scheduled automations. The preview is available on DeepSeek's official website, Pandaily reported.
Updated Sep 30, 2026 · DeepSeekDeepSeek open-sources DeepEP-Ascend and DeepGEMM-Ascend for Huawei Ascend NPUs
DeepSeek released DeepEP-Ascend and DeepGEMM-Ascend, open-source communication and GEMM libraries for Huawei Ascend NPUs. DeepEP-Ascend achieves 373–375 GB/s dispatch and 345–347 GB/s combine at EP8 on Ascend 950DT. DeepGEMM-Ascend reaches up to 99.8% of hardware limit for BF16 GEMM.
Updated Sep 30, 2026 · ByteDanceByteDance reportedly to launch personal AI product 'Spell'
ByteDance is reportedly preparing to launch a personal AI product called 'Spell', GeekPark reported, citing people familiar with the matter. The report did not specify a release date or availability outside China.
Updated Sep 29, 2026 · Beijing Academy of Artificial IntelligenceBAAI released AREX-2, a 27B self-improving agent model, on Hugging Face
BAAI released AREX-2, a 27B-parameter long-horizon agent model, on Hugging Face. It learns to improve solutions over multiple test-time rounds via propose, measure, reflect, and revise. The model is trained on machine-learning and algorithmic-programming tasks with verifiable feedback.
Updated Sep 29, 2026 · Shanghai AI LaboratoryInternLM open-sources InstantFusion, a shared latent interface for composing heterogeneous image generators
InternLM (Shanghai AI Laboratory) released the official implementation of InstantFusion, a shared latent interface for heterogeneous image generators, on GitHub. The repository includes training and inference code for latent alignment, cross-model on-policy distillation, and model handoff. Trained checkpoints are available on Hugging Face.
Updated Sep 29, 2026 · Shanghai AI LaboratoryShanghai AI Laboratory open-sources ResOPD, a sparse on-policy distillation method
Shanghai AI Laboratory released ResOPD, an open-source repository implementing Tail Residualization for Sparse On-Policy Distillation, built on verl. The repository includes a processed DeepMath subset with 1,024 training samples and 252 held-out samples.
Updated Sep 29, 2026 · Shanghai AI LaboratoryShanghai AI Laboratory open-sources AdvancedMathBench for proof generation and verification
Shanghai AI Laboratory released AdvancedMathBench, a benchmark suite with 245 proof-generation problems and 888 annotated proof trajectories, plus a trained AutoVerifier. The repository is available on GitHub and includes evaluation results for models like GPT-5.5-xhigh and DeepSeek-V4-Pro.