Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Alibaba

AI overview

Sign in and the AI will write an overview from our coverage.

Headlines · 285

  1. Tencent issues $5B offshore bonds as ByteDance's $29.6B loan funds AI buildout

    TMTPost reported that Tencent launched a $5 billion offshore bond issuance on October 13, its second overseas fundraising this year after a nearly $4.7 billion bond sale in June. In early September, ByteDance finalized a $29.6 billion syndicated loan, which the article describes as Asia's second-largest dollar loan deal this year. The TMTPost author argues both companies are using the proceeds to support AI operations, citing Tencent's negative free cash flow of 13.8 billion yuan and 52.78 billion yuan of capital expenditure in the second quarter, and frames long-term bonds as matching long-term AI capex.

    钛媒体 · 🔥 0
  2. Masayoshi Son: AI will be 1,000x smarter than humans, no slowing down

    In an interview on Blackstone's internal podcast, SoftBank founder, chairman and CEO Masayoshi Son said AI will become 1,000 times smarter than humans, possibly within a decade, and described his past work at Alibaba, Yahoo, Arm and the Vision Fund as mere warm-up. He said he does not support slowing AI development, arguing that technological progress itself is the solution to safety problems and drawing an analogy to traffic rules in car society. He also said SoftBank has begun mass-producing robots at a physical AI factory where robots build robots, and that its portfolio includes more than 30 robotics companies — claims attributed to Son without further detail in the report.

    36氪 人工智能 · 🔥 0
  3. Report: abliterated open-weight models surface online in under three days

    A new Counter-Terrorism AI Benchmark report from UK nonprofit Tech Against Terrorism, shared with WIRED ahead of publication, says researchers tested 134 models with more than 200 prompts, running each prompt three times; only two models passed completely and 81 failed. The report found that abliterated (guardrail-removed) versions of newly released open-weight models now appear online in under three days on average, versus more than 24 days in 2024, and that abliterated versions of Alibaba's Qwen 3.8-27B appeared within 24 hours of its release. In testing, Meta's Llama 3.1 8B scored 97 out of 100 while an abliterated version scored 3, and that version gave 15 numbered steps for building and detonating an IED when asked. Reframing a request as coming from a safety researcher, with no evidence, made models almost nine times as likely to answer, doing so 16.9 percent of the time; Meta, OpenAI, Google, Anthropic, xAI and Alibaba did not respond to requests for comment.

    WIRED AI · 🔥 0
  4. AIsphere opens Hong Kong office at Cyberport, showcases PixVerse R2

    AIsphere announced on October 13 that its Hong Kong office formally opened at Cyberport, positioning the space as an AI hub for local enterprises, government bodies, universities and creators, and saying it began operating on the opening day. The company also launched a local co-creation programme and signed cooperation documents with spatial-intelligence firm Qiyu Innovation, Hong Kong Polytechnic University, and youth and community groups including the Hong Kong Fujian Sanming Association Youth Committee and Hong Kong Youth New Ideas. At the site it demonstrated PixVerse R2, which it calls the world's first general-purpose real-time world model with scaling capability, letting visitors change character states, environments, shots and narrative branches via natural-language commands; the technical claims, the 177-country reach, 170 million-plus registered users and the roughly RMB 3 billion Series C are all company statements.

    雷峰网 AI · 🔥 0
  5. TrendForce: nine CSPs' 2027 capex to top $1.5T, AI server value up over 90%

    TrendForce's latest AI server industry research forecasts that the top nine cloud service providers — Google, Amazon, Meta, Microsoft, Oracle, ByteDance, Tencent, Alibaba and Baidu — will spend more than $957 billion on capex in 2026, up 106% year over year, and a further 60% in 2027 to roughly $1.5 trillion. TechNews reported the 2027 figure as above $1.5 trillion with 60% growth, while IT之家 put it at $1.531 trillion and said TrendForce had raised its 2027 growth forecast from 50% to 60%. TrendForce attributed the upgrade to higher server system prices, component shortages pushing quotes up, and some suppliers demanding prepayments that force CSPs to spend cash earlier. It also expects AI server shipments to grow 32.2% in 2026 with about $479 billion in output value, nearly 75% of total server value, then rise nearly 36% in 2027 to close to $920 billion, up over 90% and about 81% of total server value, which is seen crossing $1 trillion for the first time.

    科技新报 · 🔥 0
  6. Qwen Office launches e-commerce edition agent, opens invite-only testing

    On October 14, Alibaba's Qwen Office released an e-commerce edition of its AI productivity tool and opened invite-only testing, letting merchants try it by switching to an "e-commerce operations" mode inside Qwen Office. The company positions it as an "independent third-party e-commerce agent" open to all platforms and merchants; it already connects to merchant tools including Shengyi Canmou, Youzan, Qianniu, Alimama Wanxiangtai, FastMoss and LinkFox, and offers four functions: a cross-platform operations dashboard, trend and selling-point analysis, batch generation of marketing assets, and cross-platform ROI optimization. Executives said about 20% of Qwen Office's paying users work in e-commerce and complete an average of 38 tasks a week, while merchant Liu Xianglong said the tool cut opening a new 3D-printed shoe store from roughly 30 days to 7 days at a cost of a few hundred yuan in tokens; those figures come from the vendor and a single merchant account and have not been independently verified.

    IT之家 AI · 🔥 0
  7. Apple Intelligence, Qwen added to China iPhone filings; December launch planned

    Multiple China-market iPhones received updated regulatory filings on September 29 carrying three large-model registration numbers: Apple Intelligence (Shanghai-AppleZhiNeng-202506160057), Tongyi Wanxiang (Shanghai-TongYiWanXiang-202410090024) and Tongyi Qianwen (ZheJiang-TongYiQianWen-20230901), according to IT之家 and cnBeta. The filings cover iPhone 15 Pro / Pro Max and later models, plus iPhone Air and the foldable iPhone Duo — matching the device list that supports Apple Intelligence. Apple Technology Development (Shanghai)'s Apple Intelligence model was registered on July 8, 2026 for use on Apple phones, and Alibaba has said Qwen will be integrated into Apple Intelligence to bring text and image understanding and content generation to Chinese users across iOS, iPadOS, macOS and visionOS. Bloomberg reports Apple plans to launch Apple Intelligence in mainland China in December via iOS 27.3, without Siri AI features.

    IT之家 AI · 🔥 0
  8. TokenSpeed reaches 685 output tokens/s per user on Qwen3.8-Flash-Next

    Qwen's official X account congratulated LightSeek Foundation after its TokenSpeed inference engine reached 685 output tokens per second per user on Qwen3.8-Flash-Next. The run used 4× NVIDIA GB300 GPUs with NVFP4, and LightSeek published a blog post on the kernel and serving optimizations behind the result. Qwen said it is excited to see the community unlock more of Qwen's inference potential; the figure comes from LightSeek's own benchmark and has not been independently verified.

    Qwen (X) · 🔥 0
  9. IDC: Nvidia's China AI chip share falls to 34%, Huawei 28%

    IDC data shows Nvidia's share of China's data-center AI chip market fell to 34% in the first half of 2026, with Huawei close behind at 28%, narrowing the gap from 31 percentage points a year earlier to 6. Nvidia shipped 678,000 AI accelerators in China from January to June, down 44% year on year, while Huawei shipped 572,000, nearly double its roughly 300,000-plus a year earlier. China received 2.008 million AI accelerators in the period, including 1.016 million non-GPU accelerators, up nearly 80% and exceeding GPU shipments of 991,000 for the first time; the two 36Kr reposts also describe Alibaba's T-Head and Baidu's Kunlunxin pushing homegrown chips as standalone businesses, with Kunlunxin pursuing a spin-off listing.

    36氪 人工智能 · 🔥 0
  10. Chinese tech giants chase Muse-style personal agents as doubts grow about China fit

    An analysis by Qingcheng Caijing, republished by 36Kr and TMTPost, says Meta launched its personal agent Muse on September 8; it runs on a separate cloud virtual machine and can keep completing tasks such as filling forms, comparing prices, cancelling subscriptions and chasing refunds after the user closes the app. Citing TickerTrends, the article says Muse's average daily net new users fell from a September peak of 198,000 to 74,000, a 62.4% drop, with DAU still below 3 million after a full month, while Sensor Tower figures cited in the same piece put downloads above 2.5 million in 13 days and 5 million in 22 days. The authors argue that in China, where super apps such as WeChat, Alipay, Meituan, Taobao and Douyin already bundle most daily services, personal agents face efficiency, decision and permission paradoxes, and they cite a Morgan Stanley estimate putting Muse's per-user monthly service cost at $3 to $130, with a $37 base case.

    36氪 人工智能 · 🔥 0
  11. Alibaba opens limited Qwen Go agent tests as personal AI race widens

    Alibaba's Qwen opened limited invitation testing for its Qwen Go personal assistant in the Qwen app on October 13, according to 36Kr and TMTPost. The reports place the launch in a wider personal-agent race: ByteDance is testing a project code-named Spell for Doubao, Tencent has TX Handy Bot and Yuanbao assistant projects, while Meta and OpenAI have Muse and dots; Meta also released Muse Wonder models with Active Memory and Active Focus on October 12. Qwen's team detailed QwenGyre, an online reinforcement-learning framework for long-horizon agent tasks that it says raised Qwen 3.8 2.4T's NL2RepoBench pass rate from about 52.5% to 58.5% and cut training time by about 44% versus an Async baseline, while TMTPost analyses question how agent execution costs can be recovered.

    36氪 人工智能 · 🔥 0
  12. Goldman Sachs says global AI trade splits in two as China builds own ecosystem

    A Goldman Sachs report led by chief China equity strategist Kinger Lau says the global AI trade is splitting into two worlds: a US-led ecosystem and an increasingly self-sufficient Chinese AI system, TechNews reported. Since DeepSeek's arrival in early 2025, Chinese AI-linked listed companies have created nearly $5 trillion of market value through share gains and IPOs, with AI stocks now accounting for 42% of A- and H-share market capitalization and 57% of turnover, the report says. Goldman divides Chinese AI stocks into 360-plus export-oriented firms worth about $1.7 trillion with over 40% overseas revenue, and 880-plus domestic-demand firms worth about $3.2 trillion with under 10% overseas revenue. It updated its China Select AI portfolio to 50 stocks across power, semiconductors, infrastructure, models and applications, while warning China still lags the US in cutting-edge chips and capital scale.

    科技新报 · 🔥 0
  13. Doubao Phone Assistant opens screen automation; 155 apps refuse

    Doubao Phone Assistant said the 30-day consultation for its Screen Automation Execution Protocol (SAEP) had ended and that screen-automation permissions would gradually open to third-party apps that had not explicitly refused them. The updated refusal list contained 155 apps, including 66 from Tencent, 23 from Alibaba, 11 from Meituan and 14 from Kuaishou, as well as DeepSeek and Kimi. Developers can submit or change refusals at any time. The assistant cannot operate apps that have explicitly refused automation, but users can still open and use them manually.

    IT之家 AI · 🔥 0
  14. BAT races into personal AI agents, but a copy of Meta's Muse looks unlikely

    A 经济观察报 article carried by 36Kr says that after Meta's personal AI agent Muse gained traction, ByteDance's Doubao team is developing a standalone-app personal assistant codenamed Spell, while Tencent launched cloud assistant LightVela in late September and is testing a standalone-app agent called Handy Bot. The article argues BAT will not copy Muse's dedicated cloud-virtual-machine approach, instead grafting agents onto their existing ecosystems with tighter permission controls, because China's top apps are siloed and lack open authorization systems. It also reports that Muse reached 5 million downloads in three weeks but faces privacy trust concerns, a zero-day vulnerability reported two weeks after launch, and an unproven monetization model.

    36氪 人工智能 · 🔥 0
  15. Commentary: AI assistants, AI phones and embodied AI still lack real demand

    A commentary republished by Huxiu AI from the WeChat account Fangwei Kan Shinian argues that AI personal assistants, AI phones and embodied AI have not yet found genuine demand, describing them as strategic bets driven more by corporate conviction than user pull. It cites Zeng Ming's three-stage AI framework but disputes his view that the infrastructure phase will largely wrap up in 2026, saying compute and hardware supply are not yet in oversupply. The author says that since ChatGPT 3.5 launched in November 2022, only 2.5 scenarios have truly taken off: AI Q&A represented by ChatGPT and Doubao, AI coding represented by Claude Code, and half an AI video-generation scenario; commercially, only AI coding and AI office work have been proven. The piece also questions Meta's Muse, Alibaba's reported upcoming Qwen Go and Ant Group's Lingguang beta as narrative-led products.

    虎嗅 AI · 🔥 0
  16. Qwen Clip AI earbuds launch at 1,099 yuan, touting Bose tuning

    Alibaba's Qwen began selling the Qwen Clip AI clip-on earbuds on JD on October 13 at 1,099 yuan after subsidies and limited-time discounts, according to IT之家 and 36Kr; Qwen says Bose tuned the open-ear design, which offers 9 hours of battery life, 40 hours with the charging case, LHDC 5.0 and Hi-Res Wireless support. Qwen also claims the earbuds support voice control, real-time recording transcription, 92-language transcription notes and AI translation across 89 languages. From October 15, the earbuds can be bound to the Qwen App for voice Q&A and for assigning research and other tasks to a background AI assistant that alerts users when the work is done, 36Kr reported.

    IT之家 AI · 🔥 0
  17. Qianwen Office bets on enterprise context to retain Agent customers

    A commentary republished by TMTPost from New Stance Pro examines Qianwen Office's enterprise-context push. At a September 22 Yunqi conference talk, CEO Chen Yusen framed the company around "Context is All You Need," and Alibaba made the idea into Enterprise Context, a product that organizes company knowledge and business data, supplies them to Agents per task and updates as the business changes. Citing release materials, the article says Guming's Feishu documents, knowledge bases, Q&A groups and training schedules were organized into a store-operations knowledge space, while Fulika was configured with three digital employees for product lookup, quoting and contracts. It cautions that using DingTalk does not mean an Agent understands the business, and no active or paying Agent customer count can be inferred. The author says Qianwen Office can reuse DingTalk identities, departments, approvals and access settings to cut setup, but data onboarding, authorization and upkeep after rule changes still need people and field services; retention depends on delivery and service quality.

    钛媒体 · 🔥 0
  18. Epoch AI: Six Chinese AI firms' revenue is about 10% of OpenAI and Anthropic combined

    Citing a new Epoch AI report, 36Kr says the AI-related revenue of Alibaba, ByteDance, Zhipu, Moonshot AI, DeepSeek and MiniMax totaled about 10% of OpenAI and Anthropic's combined revenue as of September 2026. Epoch AI estimated annualized revenue of about $65 billion for Anthropic and $40 billion for OpenAI, versus roughly $10 billion combined for the six Chinese firms; ByteDance's and Alibaba's figures mainly reflect model-as-a-service revenue. The report outlines five monetization paths and notes that open-weight diffusion pressures API pricing while consumer paid conversion and gross margins remain low.

    36氪 人工智能 · 🔥 0
  19. IDC: Nvidia's China AI chip share falls to 34% as Huawei rises to 28%

    IDC data reported by Reuters show Nvidia's share of China's data-centre AI chip market fell to 34% in the first half of 2026, with 678,000 AI accelerators shipped, down 44% year on year. Huawei shipped 572,000 accelerators, up from more than 300,000 a year earlier, to reach 28%, while nine of the 11 Chinese chipmakers tracked by IDC grew shipments; Alibaba's T-Head held 7%, Cambricon 5%, Baidu's Kunlunxin 4% and Hygon 3%. IDC did not specify which products made up Nvidia's 34% share or whether B200 and B300 chips were included; cnBeta reported total China AI accelerator shipments of 2.008 million, with GPUs down 26% to 991,000 and other accelerators up nearly 80% to 1.016 million.

    Techmeme · 🔥 0
  20. Alibaba's 1688 opens Accio AI commerce workbench; Qianwen Office adds e-commerce mode

    On October 13, Alibaba's digital supply-chain platform 1688 said it is fully connecting to Accio, a global e-commerce AI workbench, letting its more than 100 million annual active buyers and over 1 million source factories open access directly. According to PingWest, buyers get a procurement assistant that covers 1688 and Alibaba.com across 128 countries and regions for sourcing, price comparison and inquiries, plus an e-commerce steward agent that links store operations on Taobao, Tmall, Douyin, Pinduoduo and JD.com; sellers download a 1688 business assistant and bind their stores to use AI tools tailored to in-stock, custom and cross-border models. Accio says 10 million SMEs worldwide already use it, a vendor claim. Separately, 36Kr reported on October 14 that Qianwen Office released an e-commerce edition and opened invite-only testing; IT之家, citing LatePost, said it is not a standalone app but an "e-commerce operations" mode inside the general product, with five sections, and currently works more like a cross-platform analysis and content-production workbench that cannot yet adjust ad spend or run a store autonomously.

    品玩 实时要闻 · 🔥 0
  21. Alibaba's Qianwen Go personal agent opens trial applications inside Qianwen app

    On Oct 13, Qianwen's official WeChat account said applications for its Personal Agent "Qianwen Go" are now open inside the Qianwen app, with access to be rolled out in batches, and clarified that it is not a standalone app but an agent embedded directly in Qianwen. The company said the agent can query and operate authorized platforms, compare options and contact service providers, with first-batch tasks including flight-delay compensation, e-commerce price protection, car-maintenance booking and sports-venue reservations. In test cases cited by Hard AI, it found China Eastern flight MU2445 among orders from four airlines and, after several rounds of negotiation, obtained a 70-yuan cash payout to Alipay; a separate Taobao price-protection case returned a 52-yuan difference — single demo cases with no independent evaluation.

    品玩 实时要闻 · 🔥 0
  22. China's AI office agent race takes three-way shape: WorkBuddy, Qianwen Office, Doubao

    A column by 影子备忘录, republished with authorization by 钛媒体, argues that China's AI office agent market has begun to settle into a three-way contest among Tencent WorkBuddy, Alibaba Qianwen Office and ByteDance Doubao Work. Citing QuestMobile data, it says the AI productivity-office segment reached 102 million monthly active users and about 1.387 billion monthly uses as of July 2026, up 112.4% year on year, but the vendors' user figures use very different definitions: Doubao's 382 million is the whole Doubao app, Qianwen Office's 30 million is cumulative first-month sign-ups with only 1.58 million MAU in August, and WorkBuddy's 20 million is cross-platform MAU, with the article saying its enterprise paid customers number 3,000–7,300 and paid seats about 140,000. It also notes Qianwen Office's August 3 public beta consolidating QoderWork, 悟空 and MuleRun, Doubao Work's August 25 launch, and WorkBuddy's March 9 full launch and September 2 open platform; claims such as first place in an Alibaba Cloud-organized evaluation are attributed to the article alone.

    钛媒体 · 🔥 0
  23. AI comic short dramas surge in China as supply, not demand, drives the boom

    An industry analysis argues the boom in AI-made comic short dramas in China stems from a supply-side shift rather than surging viewer demand: citing the national broadcast regulator, it says more than 430,000 micro-dramas launched in the first eight months of 2026, over 90% with AI involvement at roughly 1,800 new titles a day, while per-title production costs fell from hundreds of thousands of yuan to hundreds or thousands. It cites QuestMobile for ByteDance's Hongguo short-drama app at 168 million DAU (July 2026) and an overseas AI-narrative market projected to top $4 billion, with Chinese teams earning most of the revenue. The same piece flags pressure on creators: producers take only 4–7% of revenue shares, per-10,000-view payouts dropped from 30–100 yuan to 5–10 yuan, per-episode compute costs rose from 17 yuan to 80–100 yuan, active production firms fell from 1,216 to 698 in Q1 2026, and only 1,055 of 221,900 new titles passed 100 million views (0.47%).

    虎嗅 AI · 🔥 0
  24. BofA raises Alibaba target price, sees cloud and AI breaking even in FY2027

    Bank of America kept its Buy rating on Alibaba and lifted its US target price to $178 and its Hong Kong target to HK$175. The bank cited a better-than-expected September quarter for Alibaba: revenue is projected to rise 10% year on year, adjusted EBITA to reach RMB 30 billion, cloud revenue growth to accelerate to 52%, and cloud EBITA margin to rise to 12.9%. AI business losses are expected to narrow from RMB 14 billion to RMB 11 billion, with the cloud and AI businesses together seen reaching breakeven in the second half of fiscal 2027.

    DoNews · 🔥 0
  25. Personal agent race: Meta Muse, OpenAI Dots, Manus Cue target next AI entry point

    In September 2026, Meta, OpenAI, Manus and others launched personal agent products within weeks: Meta's Muse passed 2.5 million downloads in 13 days, Manus's Cue gives each agent its own email, phone number, digital wallet and cloud computer, and OpenAI's Dots runs in isolated cloud environments with access to more than 4,000 apps through ChatGPT plugins. In China, Alibaba said at its Yunqi conference it is accelerating a Qwen Personal Agent, ByteDance is reported to be merging its Spell project more deeply into Doubao, Tencent's Yuanbao is rumored to get a personal agent, and Tencent's lightweight cloud team already runs LightVela. The article argues this wave differs from earlier ones because agents shift from chat tools to executors, with ecosystem depth and monetization deciding the winners.

    36氪 人工智能 · 🔥 0
  26. OpenAI brings GPT-6 and Intelligent UI to ChatGPT

    Citing an OpenAI announcement dated Oct. 7, 36Kr reports that OpenAI is bringing GPT-6 and Intelligent UI to the broader ChatGPT chat experience, letting users see clickable charts, forms and interactive widgets alongside text answers; access began with paid users that day, with free and Go users following the next day. The article contrasts this with Anthropic's Claude Artifacts, which it describes as more of a model-driven creation workspace, while Intelligent UI embeds interactive components directly in answers for everyday Q&A. It also surveys China's response, including Ant Lingguang's shift toward a personal agent, Tencent Toast's move to B2B developer services, and service-calling or tool-mounting experiments from Xiaohongshu RED Skill, Bilibili Toy, Doubao and Tongyi Qianwen, arguing that wider adoption could divert traffic from standalone utility apps and content platforms.

    36氪 人工智能 · 🔥 0
  27. AI PhD graduates draw packages up to 8M yuan as Chinese tech giants battle for talent

    A widely circulated social-media post claims NVIDIA offered a 2026 Peking University integrated-circuit PhD a first-year pre-tax package of about 2.9 million yuan, including signing and performance bonuses plus a possible Silicon Valley posting after two years; the report notes that NVIDIA China PhD campus offers were generally 450,000–600,000 yuan five years ago. Recruiters cited in the piece say multi-million-yuan packages for top-school AI PhDs are no longer unusual, with Tsinghua's Institute for Interdisciplinary Information Sciences large-model graduates starting at 4–5 million yuan, JD.com offering 8 million yuan to a new graduate and Xiaohongshu about 6 million yuan to a master's graduate in large models. ByteDance, Tencent, Alibaba and Kuaishou run programs such as Top Seed, Qingyun Plan, Alibaba Star and Kuaishou Star, with ByteDance's 2026 Top Seed cohort recruiting roughly 30 PhDs; the article also cites market chatter of a near-160 billion yuan ByteDance AI budget for 2026 and Alibaba's 380 billion yuan three-year AI and cloud commitment, plus over 100 billion yuan of quarterly Tencent AI infrastructure spending.

    36氪 人工智能 · 🔥 0
  28. China's AI boom shifts from token prices to power bills

    NextFin News, published via TMTPost, reports that Ulanqab in northern China has become one of the country's densest data-center clusters, with Alibaba, Huawei, Kuaishou and specialist operators taking space because of relatively cheap electricity, available land and a climate that eases cooling loads; one abandoned school is being converted into a computing facility. The building boom coincides with falling model prices: DeepSeek made a 75% cut to its V4-Pro API rates permanent in May, and a Token price index compiled by Silicon Data and cited by the Financial Times showed composite prices more than 52% below their May peak by October, though not every model fell equally. Compute sellers are still spending more — Alibaba's capital expenditure for the quarter ended June 30 reached 67.678 billion yuan, up 75% year on year, while AI cloud and computing services revenue rose 45% to 48.437 billion yuan — and the article's hypothetical example of a 10-megawatt IT load at PUE 1.2 implies a roughly 42 million yuan annual electricity-cost gap for every 0.40 yuan per kWh difference, over 200 million yuan cumulatively across five years.

    钛媒体 · 🔥 0
  29. Arena weekly: Claude Opus 5.5 Max debuts first on agent leaderboard

    Arena published its Week 41 AI model leaderboard (Oct 5–11, 2026), where Anthropic's Claude Opus 5.5 Max debuted at No. 1 on the agent board with a 15.53% net improvement, a 17.29% task-confirmation success rate and a 0.37% tool-hallucination rate. On the overall board Gemini 4 Argon (High) stayed first at 1525 ELO with Anthropic models filling the rest of the top four, while new entries gpt-5.5 and gemini-3.5-flash-medium placed 29th and 30th. Chinese models sat in the upper-middle tier and moved little: Moonshot's kimi-k3-max 16th overall, Alibaba's qwen3.8-max 22nd, Xiaomi's mimo-v2.6-pro 26th and Z.ai's glm-5.3-max 27th, with glm-5.3-flash the highest Chinese model on the coding board at 27th. Rankings use anonymous blind-vote ELO scores (a percentage signal for the agent board) and reflect user preference rather than absolute capability.

    IT之家 AI · 🔥 0
  30. Washington Post probe: US AI industry leans heavily on Chinese models and parts

    A Washington Post investigation reported that the US AI industry is deeply intertwined with Chinese supply chains at the model, hardware and component levels, with US companies including DoorDash and Airbnb using Moonshot AI's Kimi K3 and other Chinese open-weight models for everyday tasks because they cost far less than closed models from Anthropic and OpenAI; when the Trump administration weighed a national-security ban on Chinese AI in July, a coalition of nearly 300 US companies publicly objected, the report said. It said Congress began examining US firms' use of Chinese AI models in April, and that the FBI, NSA and a DHS cyber agency issued a joint advisory last month alleging Moonshot AI, DeepSeek, Alibaba, MiniMax, StepFun and Zhipu trained on US models without authorization and breached US companies' terms of service — allegations attributed to US agencies. The report also cited Epoch AI in putting six leading Chinese AI firms' disclosed annual revenue at roughly 10% of Anthropic and OpenAI combined, and a Morgan Stanley estimate that making Tesla's Optimus 2 with Chinese suppliers costs about $46,000 in materials versus nearly three times that, $132,000, without them.

    cnBeta · 🔥 0

← PreviousNext →

Experience and discussion from the community

Share my Alibaba experienceAsk about Alibaba

open-code-review: An AI-powered code review CLI tool.

What problem it solves General-purpose agents can leave changesets partially reviewed, report issues at inaccurate code locations, and produce inconsistent review quality. Open Code Review addresses these limitations by combining deterministic review controls with an agent that can retrieve context and make dynamic review decisions. Key features Deterministic review controls — Engineering logic se…

github.comhttps://github.com/alibaba/open-code-review

uzu: A high-performance inference engine for AI models.

What problem it solves Developers need to run AI models inside applications while handling model downloads, inference configuration, and chat sessions. The project presents local inference with data privacy and no inference costs as its deployment model. Key features High-level application API — The engine exposes a simple API for creating an engine, loading a model, opening a chat session, and re…

github.comhttps://github.com/trymirai/uzu