Discover the AI Tools

Google just dropped LiteRT.js, a new library that makes AI run way faster in web browsers—up to three times quicker on modern hardware. It uses WebAssembly and taps into WebGPU and WebNN for hardware acceleration, leaving the old JavaScript approach in the dust. This means smoother, more powerful AI features right in your browser, no install needed.

Tencent is in talks to become the largest shareholder of AI startup Manus, with a $2 billion valuation. The deal would replace a previously planned acquisition by Meta, which has been called off. Existing investors like Tencent, Zhen Capital, and Sequoia China are backing the transaction, while US venture firms may exit. Manus will continue to operate independently and aims for a Hong Kong IPO.

Step Star is set to unveil its first AI agent phone on July 13, making it the first major large model company to enter the hardware space. The device, manufactured in deep partnership with Huqin Technology, aims to deliver a native agent experience. This move puts Step Star ahead of OpenAI, which has similar plans but not until 2027.

Samsung is quietly entering the AI PC arena with a new 4nm chip called GAIA. Designed as a dedicated NPU, it uses a memory-centric architecture to cut latency and power. Prototypes are already being tested by Lenovo and HP, with mass production expected by 2027. This move could shake up the AI PC market.

7月10日，Kimi与美国运通、中国农业银行联合推出全球首张AI原生信用卡“Kimi信用卡”。这张卡将AI服务融入核心权益，日常消费积分可兑换Kimi会员、算力额度等，实现“消费即生产力”的价值闭环。此举标志着AI大模型从软件订阅正式迈入金融生态，为智能支付和AI算力消费开辟新路径。

Samsung is making a comeback in the PC chip market with its self-developed AI accelerator, GAIA. Built on a 4nm process and centered around an NPU, the chip targets generative AI tasks. Prototypes are already with Lenovo and HP for testing, with mass production slated for 2027. Samsung also eyes robotics applications.

Meta just dropped Muse Spark 1.1, a flagship model that handles multi-agent tasks like a pro. It can juggle a main agent and sub-agents, adapt plans on the fly, and thanks to a clever context compression trick, it remembers what matters even with a million-token context window. Early tests show it crushes coding benchmarks, too.

OpenAI发布GPT-5.6系列，推出Sol、Terra、Luna三个版本，主打网络安全和性价比，同时Codex悄然下线。腾讯计划以20亿美元估值收购Manus成为最大股东，布局智能体生态。MiniMax完成160亿港元融资，创始人闫俊杰宣布零薪直至实现AGI。

Kuaishou's KwaiKAT team has released KAT-Coder-Pro V2.5, an AI model that can handle entire software projects from start to finish. Unlike typical code completion tools, this agentic model autonomously navigates codebases, fixes bugs, and runs tests. With a custom pipeline that boosts environment setup success from 16.5% to 57.2%, it's a leap toward AI that truly understands software engineering.

OpenAI 更新了 ChatGPT 的 Chrome 扩展，让 AI 助手直接嵌入浏览器侧边栏。现在，ChatGPT 不仅能回答你的问题，还能主动浏览网页、提取信息、控制标签页，甚至访问本地文件。这意味着你无需切换应用，就能在浏览器里完成更多任务。

OpenAI has unveiled the GPT-5.6 series, featuring Sol, Terra, and Luna models. Sol leads in coding benchmarks, while all models emphasize cybersecurity and cost-effectiveness. The release includes ChatGPT Work for enterprise teams, signaling a shift from parameter competition to practical value.

OpenAI has launched a new desktop application that merges ChatGPT and Codex into a single tool. The app features three modules: Chat for conversations, Work for office tasks, and Codex for programming. This move aims to simplify workflows for both developers and non-technical users, offering a unified AI assistant experience.

Lyzr, a three-year-old enterprise AI startup, just closed a $100 million Series B round—and the whole thing was handled by its own AI agent, SivaClaw. The system fielded questions from over 130 investors, drafted investment memos, and even tracked how long investors spent on slides to gauge interest. The company's valuation now sits around $500 million, and the founder never had to leave his desk for meetings on Sand Hill Road.

China Mobile has launched a new service called 'New Message Claw' that lets users remotely control desktop AI agents via SMS—no app download needed. It solves the problem of missing agent notifications when away from your computer. Compatible with several open-source agents, it uses operator SMS channels for low-latency, reliable control.

OpenAI's second-in-command, Fidji Simo, is stepping down from her full-time role due to a recurring neuroimmune disease. She'll stay on as a part-time consultant. The departure comes at a critical time as OpenAI faces slowing ChatGPT growth and prepares for a potential IPO, putting pressure on CEO Sam Altman to find a successor quickly.

Google has revamped its Android Bench code ranking, introducing a secure Harbor sandbox framework for fairer testing. The new benchmark sees Anthropic's Claude5 top the list with 84.5% accuracy, while Google's own Gemini models struggle—Gemini 3.5 Flash took 28 hours and $165 to run. The move highlights the industry's shift toward autonomous development workflows and positions Android Bench as an authoritative AI evaluation platform.

Apple is reportedly in talks with AI startup PrismML to bring large language models directly to iPhones. Their 1-bit compression tech shrinks model size to 1/14, boosts speed 8x, and cuts power use by 80%. A 27B-parameter Qwen model already runs on an iPhone 17 Pro, hinting at a major shift in on-device AI.

Meta is teaming up with Broadcom and TSMC to produce its own AI chip, Iris, set for mass production in September 2026. The chip aims to optimize ranking and recommendation systems on Facebook and Instagram, reducing reliance on third-party GPUs and cutting data center costs. Meta also plans to double its computing capacity by 2027.

Anthropic has launched a test version of its Reflect feature, turning Claude into a personal AI usage coach. It analyzes your chat history across four dimensions—task delegation, goal description, result identification, and responsible caution—to generate a personalized 'health check' report. Available to free, Pro, and Max users with memory enabled, Reflect aims to improve your AI collaboration skills while respecting privacy by excluding incognito chats and health data.

OpenAI launched GPT-5.6 on July 9, designating it as the preferred model for Microsoft 365 Copilot across Word, Excel, PowerPoint, and Teams. The move counters recent speculation that Microsoft might replace OpenAI's models with its own MAI system to cut costs. While neither company disclosed full partnership details, the announcement reinforces their strategic alliance amid fierce competition from xAI's Grok4.5 and Meta's Muse Spark1.1.