Discover the AI Tools

豆包近期对其任务模式进行了重大更新，推出了技能、连接器和工作伙伴三大新功能，标志着其从聊天助手向智能办公平台的转变。目前平台已拥有超过200种技能和连接器，用户还能自定义技能，打造标准化工作流。这些功能可单独或组合使用，旨在简化复杂工作流程，提升办公效率。

SAP has joined forces with esports powerhouse Team Liquid to create an AI voice application that turns chaotic team chatter into structured match data. By separating individual voice streams, transcribing speech, and analyzing emotions, the tool helps teams review strategies and improve performance. This partnership signals a shift from simple transcription to specialized analytics in gaming.

SenseTime has open-sourced SenseNova U1.5Lite, an 8B-parameter multimodal model that punches above its weight. It handles 4K output natively and excels at complex visual editing tasks, rivaling much larger commercial models. This lightweight powerhouse is set to democratize advanced AI image generation and editing for developers and creators alike.

Liquid AI and Hugging Face have released DSpark draft models for the LFM2.5 series, boosting inference speed up to 3.18x on GPU and 2.87x on edge devices, with no quality loss. The technology uses speculative decoding to cut latency, making local AI agents faster than some cloud models. Here's how it works and what it means for developers.

Mistral AI has launched Agentic Search, an upgraded retrieval system that tackles the limitations of traditional RAG. By using a multi-step loop with tools like search, open, navigate, read, and grep, it can reason and iterate, boosting accuracy on complex documents to 86% in tests. This breakthrough helps enterprises unlock insights from lengthy financial reports, legal contracts, and government records.

Alibaba has open-sourced Qwen-UI-Agent, a multimodal foundation model designed for real-world GUI interaction. It excels across mobile, desktop, and web environments, outperforming industry leaders in benchmarks like MobileWorld and OSWorld. The model's ability to operate on real devices, coupled with robust safety mechanisms, marks a significant step toward practical AI assistants.

OpenAI has taken a significant step in democratizing AI coding by fully open-sourcing Codex Harness, the core runtime that powers its Codex agent. This move allows developers to build custom AI programming tools and workflows, extending competition from model capabilities to the entire agent execution framework. With Harness, developers can now integrate Codex's intelligence into their own IDEs, apps, and more, marking a shift from mere code generation to a complete agent-based development paradigm.

Kakao's board has approved a plan to split the company, creating a new entity called Kakao AI that will focus on KakaoTalk and artificial intelligence. The remaining business, to be renamed Kakao X, will handle investments, fintech, content, and mobility. This move aims to sharpen the AI strategy and boost capital efficiency, with the split expected to complete by early 2027.

At its semi-annual earnings briefing, iFLYTEK announced a major upgrade: a new flagship general-purpose large model, built entirely on domestic computing power, will debut at the 1024 Developer Festival. A phased version arrives by end of August. This model promises top-tier code capability and cost-effectiveness, marking a significant stride for China's AI self-reliance.

TaoTie has launched its iO AI glasses at 2499 yuan, with a promotional price of 1996 yuan. These 'human enhancement' glasses feature a dual-eye display, 24/7 AI memory, and privacy-first design. Available on major Chinese e-commerce platforms, they support prescriptions up to 1000 degrees and come with a two-day battery life.

DeepSeek has quietly rolled out a significant update to its developer toolchain, introducing a new visual model experience that brings multimodal capabilities to the table. The V4Flash Vision-Exp model, part of the DeepSeek Harness 0.1.1-rc.1 release, allows AI agents to process both text and images, making it easier for developers to build smarter applications. This update marks a major step forward in integrating vision into everyday coding workflows.

ByteDance's Seed division has reorganized, creating four new departments to streamline AI development. The move separates enterprise-focused agent models from consumer chat models, with a renewed emphasis on reinforcement learning. This restructuring aims to boost model intelligence and tailor products for different users, signaling a strategic shift in the company's AI approach.

Google DeepMind's open-source Gemma models have surpassed one billion downloads, spawning over 100,000 variations. The models are now powering space missions with NASA, aiding medical diagnostics in India and Uganda, and even decoding dolphin communication. This milestone highlights Gemma's versatility across extreme environments and its growing impact on real-world applications.

OpenAI has rolled out a significant update to ChatGPT's macOS app, enabling deep integration with Apple's Messages. With user permission, the AI can read, compose, and send texts, and even summarize chat histories. This move has caught the attention of Bloomberg's Mark Gurman, who sees it as a direct challenge to Siri's ecosystem advantages. The feature requires several permissions, including full disk access, and could signal a new era of AI-powered communication.

OpenAI has regained the lead in the U.S. enterprise AI market, surpassing Anthropic with over 40% share. Ramp's data shows shifting dynamics as spending on AI tools continues to climb. What's driving this change? New models, pricing strategies, and data policies are reshaping customer choices.

xAI旗下的聊天机器人Grok近日出现异常，部分用户在Grok.com上请求生成PDF后，收到了连续的无意义乱码文本，甚至包含强化学习研究相关的链接。问题主要影响Grok Lite用户，刷新或重启会话通常能恢复正常。官方回应称这是“罕见的临时生成故障”，并建议用户重启对话。此次事件发生在xAI人员流失严重的背景下，再次凸显了大模型在推理、生成和检索链路上的稳定性挑战。

Alibaba has unveiled Qwen-UI-Agent, a new GUI agent model that excels across mobile, desktop, and web environments. In head-to-head tests, it outperformed GPT-5.6Sol and Claude Opus4.8 on mobile benchmarks, scoring 82.1% on MobileWorld and 97.5% on AndroidDaily. The model also supports real-device training and robust security features, marking a significant leap in GUI automation.

在2026世界机器人大会上，宇树科技创始人王兴兴透露，公司正推进物理AI机器人模型的自我进化，并探索用大模型自动生成机器人控制代码。他认为，具身智能的“ChatGPT时刻”仍需突破泛化能力瓶颈，预计2到3年内可能迎来转折点。

Tencent's next-gen AI model, Hy4, has surfaced in internal tests, hinting at an imminent release. Positioned as an expert-level model, it's set to power the WorkBuddy assistant. With potential upgrades like trillion-scale parameters or advanced reinforcement learning, Hy4 aims to tackle complex tasks, building on the success of its predecessor Hy3.

Apple Music is making AI transparency labels mandatory for AI-generated tracks by late 2026, a shift from voluntary disclosure. The move is part of broader efforts to curb AI spam, including internal monitoring tools and royalty reallocation. With over a third of uploads now AI-generated but barely listened to, platforms like Deezer and Spotify are also tightening their rules, signaling a new era of accountability in music.