中文 ↗RSS ↗

THE NEWSROOM / AI 新闻

AI News · Page 12

Source-linked briefings on models, agents, products and practical AI.

YYSI publication dates appear on the left; original announcement dates appear under each brief.

Search and categories filter this page. Use the page links below for earlier news.

SOURCE

Practical guides Google

Google launches AI certificate for healthcare professionals

Google announced the Google AI for Healthcare Professionals Certificate, a self-paced program teaching responsible, practical AI use for non-clinical administrative work. Google says it takes under five hours and offers up to 5.0 AMA PRA Category 1 Credits. The company states NYU Grossman School of Medicine, University of Alabama at Birmingham, Touro University and Purdue University School of Nursing are offering it to faculty, residents, students and nursing trainees.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Models & research Mistral

Mistral releases open weights for Mistral Large 4 (Le Chonk)

Mistral has released the weights of Mistral Large 4, codenamed Le Chonk, a 1-trillion-parameter natively multimodal model first shown in public preview earlier this month. Mistral says additional reinforcement learning improved agentic coding and defensive cybersecurity, and that the model leads open-weight coding benchmarks while using fewer tokens. It adds three reasoning modes and is trained on European infrastructure.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Google

Google partners with Singapore to deploy frontier AI across public health

Google announced at its Singapore health summit a deepened partnership with Singapore to deploy frontier AI across the national healthcare system. Plans include a new Google Research hub in Singapore focused on foundational AI models, expanded use of open models like MedGemma in frontline care, a $7 million contribution from Google.org and Google DeepMind to the Philanthropy Asia Alliance, and DeepMind's evaluation of an AI co-clinician tool with the National University Health System.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Google

Google expands ads in AI Mode to ten new countries

Google said it is scaling ads in AI Mode to Australia, Canada, India, Kenya, Malaysia, New Zealand, Nigeria, Pakistan, the Philippines and Singapore, with plans to add Brazil, Mexico and the UK in Q4 2026. The company also launched multiturn ads in the U.S. for follow-up queries, labeled "Sponsored." Google claims people find AI Mode ads relevant and that click-through rates are similar to organic product list items.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Google

Google offers Canadian students one year of Gemini AI Plus free

Google Canada announced that eligible university and college students in Canada can claim one year of Google AI Plus at no cost, auto-renewing at $6.99 per month unless cancelled. The offer includes doubled Gemini usage limits, Gemini in apps like Gmail, 400GB of storage, and a new student hub with study notebooks, flashcards, practice quizzes and interactive visualizations. The company says the plan must be claimed by December 31.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Google

Gemini app adds library of immersive explorations for students

Google says the Gemini app now offers a library of immersive explorations, an interactive, visual way to explore topics layer by layer, from microscopic systems to deep space. The library is reachable through the Gemini student hub via a "Try immersive explorations" option, with direct access in chat conversations coming soon. Google says audio narration is also planned, and the feature is available to all users in English, Indonesian, Hindi, Brazilian Portuguese, French and simplified Chinese.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Google

Chrome enhanced autofill pulls form suggestions from Gmail via Gemini

Google says Chrome's enhanced autofill can now suggest details pulled directly from Gmail while users fill out online forms, powered by Gemini with Personal Intelligence. The company claims it can surface order numbers, tracking information and flight bookings. Google says it plans to extend the feature to images in Google Photos, and that the capabilities are beginning to roll out in the U.S. on desktop, Android and iOS for eligible Gemini in Chrome users.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Models & research Google

Google announces Gemini 4 Carbon frontier model with cybersecurity focus

Google announced Gemini 4 Carbon, a frontier model it says leads on cybersecurity and long-horizon agentic tasks. Google claims Carbon ranks first on the Artificial Analysis Intelligence Index v5.0 at 59 and scores 88.2% on SEC-bench Pro, 54.8% on ExploitGym and 71.0% on CWE-bench. Access begins with select cyber defenders via the Fairwind Program; pricing matches Argon at $2/$10 per million input/output tokens.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Google

Google and XREAL launch first wired Android XR glasses, XREAL AURA

Google and XREAL launched XREAL AURA, which they describe as the first wired XR glasses powered by Android XR. The glasses weigh under 95 grams, with processing handled by a tethered compute puck carrying the battery and Qualcomm's new Snapdragon Reality Elite platform. Google says Gemini can understand what users see and do, and the device supports multi-app workspaces, wired laptop mirroring, immersive video and games. It is available to order in the U.S., Canada and Korea, with launch pricing from USD 1,099.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Models & research Google

Google unveils Genie 4 world model with native audio and real-time control

Google announced Genie 4, a world model that generates interactive virtual worlds in real time, including native ambient sound and character speech. Google says it integrates Gemini models, with Gemini 3.8 Flash acting as a guiding agent for multi-step events. Genie 4 is in a research preview with trusted testers, with planned expansion to Google AI Ultra subscribers via Project Genie and early-access partners on the Gemini API and Google AI Studio.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Agents & tools Meta

Meta launches Muse personal AI agent in Singapore

Meta says its Muse personal AI agent is now available in Singapore, following its earlier global announcement. The agent connects to services such as Gmail, Google Calendar and Instagram, plus local partners including Shopee, Grab, Singapore Airlines, Agoda, ShopBack, Klook, Trip.com and Lazada. Meta says Muse runs on a dedicated cloud computer called Muse Secure VM and asks users before sensitive actions like purchases or sending email. It is available on iOS, Android and muse.ai, with AI glasses coming soon.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Anthropic

Anthropic Expands AI for Science Program to Foundations and Research Institutions

Anthropic says it is expanding its AI for Science program, which provides free Claude credits to researchers, to partner with 20 scientific foundations, institutions and research groups worldwide. Partners include the Sloan and Simons foundations, Schmidt Sciences, the US National Science Foundation, the Wellcome Sanger Institute and the Francis Crick Institute. Anthropic also cites work with HHMI Janelia on a lab-instrument standard and with the Allen Institute on AI agent teams.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Kimi

Kimi Code Desktop launches as native app for AI coding agents

Kimi has released Kimi Code Desktop, the official desktop version of Kimi Code, which brings local projects, agent conversations, code changes, terminal checks and browser verification into one graphical workspace. The vendor says users can ask the agent to read and write code, run commands and automate tasks, while reviewing diffs and command output. It supports macOS, Windows and Linux, and works with a Kimi account or third-party model providers.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Anthropic

Anthropic Expands GSA Partnership With 40% Claude Discount for Government

Anthropic said it expanded its General Services Administration partnership with a new agreement giving government buyers a 40% discount on Claude Enterprise, Claude for Government and the Claude API, extending to state, local, tribal and territorial governments. Anthropic estimates the discounts will save the government $1 billion, with no minimum spend commitment. The company also launched a program to help agencies implement Claude, and said roughly 500 US government organizations now use it.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Anthropic

Anthropic Releases Claude for Nonprofits Plugin

Anthropic announced the Claude for Nonprofits plugin, a set of skills and connectors developed with roughly 30 nonprofit design partners. According to the company, it helps organizations configure Claude with their own context and policies, supports fundraising from research through reporting, and handles operational tasks for small teams. Anthropic also said it plans to extend its Claude Corps fellowship program through a partnership with Taproot.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Anthropic

Anthropic files S-1 for proposed IPO on Nasdaq

Anthropic, PBC announced it has filed a Form S-1 registration statement with the U.S. Securities and Exchange Commission for a proposed initial public offering of its Class A common stock. The company intends to list on the Nasdaq Global Select Market under the symbol "ANT." The number of shares to be offered and the price range have not yet been determined, and the registration statement has not yet become effective.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Models & research Google

Google launches Gemini 4 Argon frontier model

Google says Gemini 4 Argon, its new frontier model, is built for complex, multi-step work across coding, finance, legal, cybersecurity and science, and that it outperforms other frontier models on cross-domain benchmarks. The company says Argon is now globally available to developers and enterprises through the API in AI Studio and the Gemini platform, and to subscribers in Antigravity, the Gemini app and AI Mode in Search.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates Google

Google and Voltus expand to a 1-gigawatt virtual power plant

Google and Voltus announced an expanded agreement to build a 1-gigawatt virtual power plant across the PJM grid region, which serves 67 million people in 13 states and Washington, D.C. The plant aggregates flexible capacity from businesses and homes, such as smart thermostats, batteries and industrial assets, to reduce energy use during peak demand. It builds on an earlier 100-megawatt agreement, a tenfold expansion.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Product updates MiniMax

MiniMax Code launches Mini App runtime for persistent agent workspaces

MiniMax announced Mini App, a new extension format for its MiniMax Code desktop app that packages a frontend and Node.js backend as an installable plugin. According to the company, Mini Apps keep agent work on a live page rather than scattered across conversations, and can call and be called by the Agent. The release includes a runtime, a /miniapp-creator flow, and official Mini Apps in the plugin marketplace.

AI summary · Official sourceOriginal source ↗Read brief →
SOURCE

Practical guides Anthropic

Anthropic Opens Call for Life Sciences Guest Investigators

Anthropic announced an open call for principal investigators and research groups at academic institutions, nonprofits, and government labs to apply for 8-12 week guest investigator projects using Claude. The company says it will run its models and harnesses at no cost, collaborate on publishing outputs, and open-source accompanying data and code. Applications are open for roughly 14 days, with selections notified about two weeks after closing.

AI summary · Official sourceOriginal source ↗Read brief →