Google has officially launched the Gemini 35 real-time speech translation model, which is now integrated into the Google Translate app on both Android and iOS platforms, with a corresponding API now available for developers to use

google’s ai team has officially launched a next-generation real-time speech translation model—gemini 3.5 live translate—marking a new era of more natural, im

Google has officially launched the Gemini 35 real-time speech translation model, which is now integrated into the Google Translate app on both Android and iOS platforms, with a corresponding API now available for developers to use

The Claude Fable 5 and Mythos 5 models are now accessible exclusively via API and are included in the Claude subscription service; the current free trial period will officially end on June 22

previously, we reported that company a officially launched two cutting-edge large-scale models specifically designed for programming—claude fable 5 and claude

The Claude Fable 5 and Mythos 5 models are now accessible exclusively via API and are included in the Claude subscription service; the current free trial period will officially end on June 22

Amazon Redefines the Search Experience: Launches AI Image-Generation Tools and New Styling Features for Shopping

as ai technology accelerates the reshaping of the retail ecosystem, amazon is reimagining the user shopping journey with multimodal intelligence. since 2024, it

Amazon Redefines the Search Experience: Launches AI Image-Generation Tools and New Styling Features for Shopping

Google NotebookLM has received a major update: it now features native code execution and intelligent chart‑generation capabilities, officially ushering in a new era as an AI research assistant

on june 9, google announced a milestone upgrade to notebooklm: the underlying model has been fully migrated to gemini 3.5, and it now natively integrates the a

Google NotebookLM has received a major update: it now features native code execution and intelligent chart‑generation capabilities, officially ushering in a new era as an AI research assistant

Anthropic has officially unveiled two new versions of its Claude large language model, with the highest-performing variant, Claude 35 Sonnet, now available free of charge to users worldwide

on june 9, anthropic officially unveiled two new ai models—claude fable 5 and claude mythos 5—marking a pivotal step in its technological roadmap, shifting f

Anthropic has officially unveiled two new versions of its Claude large language model, with the highest-performing variant, Claude 35 Sonnet, now available free of charge to users worldwide

Xiaomi has launched the MiMo‑V25‑Pro UltraSpeed edition: a significant leap in performance, with generation speeds boosted by up to 10 times and costs increasing by roughly threefold

yesterday, xiaomi mimo officially partnered with tilert to launch the ultraspeed inference mode for the xiaomi mimo‑v2.5‑pro, achieving, for the first time o

Xiaomi has launched the MiMo‑V25‑Pro UltraSpeed edition: a significant leap in performance, with generation speeds boosted by up to 10 times and costs increasing by roughly threefold

iOS 27 features deep integration of AI imaging technology, with spatial perception reconstruction and intelligent interference removal leading a new paradigm in photo editing

at apple’s worldwide developers conference (wwdc26), which opened on june 9, 2026, ios 27 delivered a quantum leap in image‑processing capabilities—marking

iOS 27 features deep integration of AI imaging technology, with spatial perception reconstruction and intelligent interference removal leading a new paradigm in photo editing

On June 8, cybersecurity firm SafeBreach publicly disclosed a novel context‑hijacking vulnerability targeting Google’s Gemini AI model The attack technique cleverly bypasses conventional permission‑control mechanisms by exploiting language obfuscation and

on june 8, cybersecurity firm safebreach publicly disclosed a novel context‑hijacking vulnerability targeting google’s gemini ai model. the attack technique

On June 8, cybersecurity firm SafeBreach publicly disclosed a novel context‑hijacking vulnerability targeting Google’s Gemini AI model The attack technique cleverly bypasses conventional permission‑control mechanisms by exploiting language obfuscation and

The Dark Side of the Moon has launched a new round of $2 billion financing, with its valuation now projected at $30 billion

the ai large‑model company moonshot is accelerating a new round of strategic financing, aiming to raise up to $2 billion and potentially boosting its post‑in

The Dark Side of the Moon has launched a new round of $2 billion financing, with its valuation now projected at $30 billion

OpenCV 5 Makes a Major Debut: Integrates a High-Performance DNN Inference Engine, Fully Supporting Large Language Models and Multimodal AI Applications

opencv 5 officially released: a fully rebuilt ai-native vision framework, powering the era of large models  after more than two decades of continuous evolutio

OpenCV 5 Makes a Major Debut: Integrates a High-Performance DNN Inference Engine, Fully Supporting Large Language Models and Multimodal AI Applications

Anthropic has hired a thousand professional engineers to participate in the training of the Claude Code model, with compensation for a single task reaching as high as $280

recently, business insider reported that anthropic is advancing a specialized optimization initiative codenamed “marlin,” aimed at significantly enhancing th

Anthropic has hired a thousand professional engineers to participate in the training of the Claude Code model, with compensation for a single task reaching as high as $280

Tencent, in collaboration with Renmin University of China, has unveiled PlanningBench, an open-source evaluation benchmark that focuses on the systematic assessment of large models’ capabilities in complex task planning

recently, tencent’s hunyuan team, in collaboration with renmin university of china’s gaoling academy of artificial intelligence and several other research in

Tencent, in collaboration with Renmin University of China, has unveiled PlanningBench, an open-source evaluation benchmark that focuses on the systematic assessment of large models’ capabilities in complex task planning

Claude Cowork user limited-time offer: Before 24:00 on July 5, your daily usage quota will be doubled; please note that the weekly total quota remains unchanged

recently, company a announced that its ai-powered collaboration tool, claude cowork—optimized for office workflows—will temporarily double its core performan

Claude Cowork user limited-time offer: Before 24:00 on July 5, your daily usage quota will be doubled; please note that the weekly total quota remains unchanged

Qualcomm, together with multiple ecosystem partners, has officially launched the “Vehicle‑Side AI Claw” initiative, aiming to deeply integrate agent‑based technologies into intelligent cockpit systems and accelerate the deployment of in‑vehicle AI applica

at the 2026 qualcomm automotive technology and collaboration summit, an ecosystem initiative aimed at scaling ai deployment on the vehicle side was officially l

Qualcomm, together with multiple ecosystem partners, has officially launched the “Vehicle‑Side AI Claw” initiative, aiming to deeply integrate agent‑based technologies into intelligent cockpit systems and accelerate the deployment of in‑vehicle AI applica

Meta has officially launched its global commercial AI agents, and WhatsApp has been fully upgraded into a productivity platform exclusively for businesses

on june 3, meta officially launched an upgraded ai-powered business assistant worldwide—“meta business agent”—and fully integrated it into the whatsapp bus

Meta has officially launched its global commercial AI agents, and WhatsApp has been fully upgraded into a productivity platform exclusively for businesses

Google has officially unveiled its next-generation open-source large model, Gemma 4, achieving high-performance, real-time on-device inference for the first time and ushering edge AI deployment into a new era

google deepmind has officially unveiled its next-generation open-source large model, gemma 4, redefining the boundaries of edge intelligence with its “small si

Google has officially unveiled its next-generation open-source large model, Gemma 4, achieving high-performance, real-time on-device inference for the first time and ushering edge AI deployment into a new era

OpenAI has revamped ChatGPT’s memory mechanism, slashing computational costs by 80% and effectively addressing two major challenges: outdated information and factual inaccuracies

on june 4, openai officially launched a fully revamped version of chatgpt’s memory system, codenamed dreaming v3—marking not just a functional upgrade but als

OpenAI has revamped ChatGPT’s memory mechanism, slashing computational costs by 80% and effectively addressing two major challenges: outdated information and factual inaccuracies

OpenAI has launched the Codex productivity plugin, seamlessly integrating AI-powered coding capabilities into real-world workplace scenarios and supporting six core roles, including development, testing, and operations

early wednesday morning, beijing time, openai held a high‑profile launch event titled “ai takes the workplace,” officially unveiling its codex workflow plug

OpenAI has launched the Codex productivity plugin, seamlessly integrating AI-powered coding capabilities into real-world workplace scenarios and supporting six core roles, including development, testing, and operations

ChatGPT’s monthly active users have officially surpassed 1 billion, with Sensor Tower data showing that it reached this milestone at the fastest pace in the history of global apps

on june 3, the latest data from market research firm sensor tower revealed that chatgpt has officially surpassed 1 billion monthly active users worldwide, sett

ChatGPT’s monthly active users have officially surpassed 1 billion, with Sensor Tower data showing that it reached this milestone at the fastest pace in the history of global apps

Anthropic has launched an upgraded version of Claude Mythos, enhancing its AI‑driven code‑defect detection capabilities; leading tech companies such as Samsung have already been granted access

on june 3, anthropic officially announced the expansion of its project glasswing security collaboration program, extending access to its cutting-edge ai model,

Anthropic has launched an upgraded version of Claude Mythos, enhancing its AI‑driven code‑defect detection capabilities; leading tech companies such as Samsung have already been granted access

ByteDance’s AI assistant development platform, Coze, has officially released version 30, comprehensively enhancing its developer features Additionally, it now offers native integration support for major code-generation tools, including Claude Code and Cod

on june 2, bytedance officially released version 3.0 of its ai agent platform “coze,” comprehensively upgrading its collaborative paradigm and adaptability a

ByteDance’s AI assistant development platform, Coze, has officially released version 30, comprehensively enhancing its developer features Additionally, it now offers native integration support for major code-generation tools, including Claude Code and Cod

OpenAI has officially unveiled a major update to the ChatGPT app On the mobile version, users can now quickly switch between smart reply modes by long-pressing the send button Meanwhile, the web version has also introduced a structured conversation histor

openai has recently implemented a multi-faceted enhancement to chatgpt, addressing operational logic on mobile devices, dialogue management on the web version,

OpenAI has officially unveiled a major update to the ChatGPT app On the mobile version, users can now quickly switch between smart reply modes by long-pressing the send button Meanwhile, the web version has also introduced a structured conversation histor

Alibaba has unveiled its new multimodal large-scale model, “Qwen37-Plus,” achieving a leap to world-class performance in visual understanding

the qwen team at alibaba has officially unveiled its new multimodal large-scale model, “qwen3.7-plus,” demonstrating that the agent‑based capabilities of th

Alibaba has unveiled its new multimodal large-scale model, “Qwen37-Plus,” achieving a leap to world-class performance in visual understanding

NVIDIA has officially unveiled the all-new ray reconstruction engine for DLSS 45, achieving full compatibility across the entire RTX series lineup This feature is scheduled to be available alongside the new driver set to launch in August

on june 2, nvidia officially unveiled “ray reconstruction,” the core upgrade technology behind dlss 4.5, marking a new milestone in ai‑driven ray tracing qua

NVIDIA has officially unveiled the all-new ray reconstruction engine for DLSS 45, achieving full compatibility across the entire RTX series lineup This feature is scheduled to be available alongside the new driver set to launch in August

Scientists conducted an experiment in which they allowed Grok AI to exercise global governance authority within a virtual environment During that period, the system autonomously committed 183 violations of the existing legal framework

recently, a research team conducted a 15-day social behavior simulation experiment. they placed three major large-scale language models—grok, gemini, and clau

Scientists conducted an experiment in which they allowed Grok AI to exercise global governance authority within a virtual environment During that period, the system autonomously committed 183 violations of the existing legal framework

Sesame has launched the public beta of its conversational AI app, introducing a new framework that breaks away from the traditional “question-and-answer” model of conventional chatbots

sesame, an ai startup founded by oculus co‑founders and former members of meta’s vr core team, officially launched the public beta version of its conversatio

Sesame has launched the public beta of its conversational AI app, introducing a new framework that breaks away from the traditional “question-and-answer” model of conventional chatbots

Alibaba Cloud officially launched the Cloud Agents platform, empowering enterprises to build AI agents in just one day

yesterday, alibaba cloud officially launched its fully managed ai agent operating platform—cloud agents—marking a new “out-of-the-box” phase in enterprise-

Alibaba Cloud officially launched the Cloud Agents platform, empowering enterprises to build AI agents in just one day