Core Insights - LongCat-Flash-Thinking enhances the autonomous tool-calling capabilities of agents and expands formal theorem proving abilities, becoming the first domestic large language model with both "deep thinking + tool calling" and "non-formal + formal" reasoning capabilities [3] - The new model shows significant advantages in handling high-complexity tasks such as mathematics, coding, and agent tasks [3] - LongCat-Flash-Thinking is fully open-sourced on HuggingFace and GitHub, allowing users to experience it on the official website [3]

【环球网科技报道 记者 李文瑶】9月22日，美团宣布高效推理模型 LongCat-Flash-Thinking正式发布。据介绍，新模型除保持龙猫模型一贯"快"的特点的同 时，在逻辑、数学、代码、智能体等多个领域的推理任务中，也达到了全球开源模型的最先进水平（SOTA），部分任务性能接近闭源模型GPT5-Thinking。 目前，LongCat-Flash-Thinking已在HuggingFace、Github全面开源，用户可在官网体验。 同时，LongCat-Flash-Thinking增强了智能体自主调用工具的能力，并扩展了形式化定理证明能力，成为国内首个同时具备"深度思考+工具调用"与"非形式化 +形式化"推理能力的大语言模型。该团队还表示，尤其在高复杂度的任务（如数学、代码、智能体任务）处理上，新模型具备显著优势。 ...

Group 1 - The article discusses the initial skepticism towards AI advancements due to many major companies' new versions falling short of expectations, leading to concerns about future developments [1] - However, after participating in discussions about AI implementation, there is a renewed optimism as AI is being widely adopted across various industries, subtly transforming the world [2] - There is a distinction made between high-end AI technologies, such as large language models and autonomous driving, and more mature technologies like voice and image recognition, which are not considered groundbreaking [3][4] Group 2 - AI is portrayed as a tool accessible to everyone, capable of improving efficiency and outcomes in specific scenarios, thus demonstrating the value of technological progress [4][5] - Numerous practical applications of AI are highlighted, such as automatic transcription of meetings and structured processing of customer interactions, which significantly enhance digital capabilities [6][7] - Despite some professionals dismissing these applications, they are recognized as valuable and memorable, showcasing AI's ability to meet user needs and gain acceptance [8] Group 3 - The article notes that while AI is a hot topic in the tech industry, many projects are still struggling to achieve profitability, indicating that the AI industry is not yet stable [23][24] - From a supply-side perspective, AI applications often lack economies of scale, as the backend systems require extensive customization, making it difficult to standardize and productize solutions [25] - Users expect AI applications to be cost-effective, and while many AI solutions are currently subsidized, the sustainability of this model is questioned as companies may eventually need to charge for their services [27] Group 4 - There are differing opinions on the future of AI, with some focusing on continuous investment in generative AI, while others seek to commercialize existing technologies and create tangible value [28][31] - The article suggests that without a consensus in the industry, the fragmented approach to investment may lead to suboptimal outcomes for all parties involved [32] - Despite the challenges, the gradual integration of AI is enhancing the overall digital landscape, benefiting both individuals and organizations [33]

官方介绍，该模型不仅增强了智能体自主调用工具的能力，还扩展了形式化定理证明能力，成为国内首 个同时具备"深度思考+工具调用"与"非形式化+形式化"推理能力相结合的大语言模型。尤其在超高复杂 度的任务(如数学、代码、智能体任务)处理上，LongCat-Flash-Thinking具备更显著的优势。 智通财经APP获悉，9月22日，美团(03690)发布高效推理模型LongCat-Flash-Thinking。美团表示，基于 AIME25实测数据，LongCat-Flash-Thinking在该框架下展现出更高效的智能体工具调用能力，在确保 90%准确率的前提下，相较于不使用工具调用节省了64.5%的Tokens。目前，该模型已在HuggingFace、 Github全面开源。 综合评估显示，LongCat-Flash-Thinking在逻辑、数学、代码、智能体等多个领域的推理任务中，达到了 全球开源模型的最先进水平(SOTA)。 ...

Core Insights - Meituan has officially released its efficient reasoning model LongCat-Flash-Thinking, which maintains the speed characteristic of its predecessor, the LongCat model, while achieving state-of-the-art (SOTA) performance in reasoning tasks across various domains such as logic, mathematics, code, and intelligent agents, with some tasks nearing the performance of the closed-source model GPT5-Thinking [1] - The LongCat-Flash-Thinking model enhances the autonomous tool-calling capabilities of intelligent agents and expands its formal theorem proving abilities, making it the first domestic large language model to combine "deep thinking + tool calling" with "non-formal + formal" reasoning capabilities [1] - The new model demonstrates significant advantages in handling high-complexity tasks, particularly in mathematics, code, and intelligent agent tasks [1] - LongCat-Flash-Thinking is fully open-sourced on platforms like HuggingFace and GitHub, and it is available for experience on the official website [1]

Group 1 - The consumer electronics sector is experiencing a peak production period with a concentration of new product launches from September to October [4] - Semiconductor stocks continue to show strong performance, with companies like Demingli and Wanrun Technology hitting their limits [1] - The stock of Taimusi experienced a significant drop after a period of rapid gains, indicating volatility in the market [1] Group 2 - The consumer electronics sector has potential for rebound, with companies like Luxshare Precision and Heertai seeing significant stock price increases [3] - OpenAI's collaboration with Luxshare Precision to develop a revolutionary AI device is expected to create new market opportunities [3] - The shift of AI trends from cloud to edge devices is seen as a critical development, potentially leading to broader opportunities in edge devices, computing chips, and communication modules [4] Group 3 - A total of 13 consumer electronics stocks have doubled in price this year, with notable increases from companies like Chipone and Industrial Fulian [5] - Over 30 consumer electronics stocks have received institutional research attention, indicating heightened market interest [5] - Companies like Celeritek and Dongshan Precision are expected to benefit from the growing demand for AI computing, with projections of continued high growth in their earnings [6]

Gemini和Cla ude 还在冲！ 如上图，chatgpt 日活走平了！ - Standard_NV18ads_A10_v5 Standard_NV36adms_A10_v5 - Standard_NV12ads_A10_v5 = - Standard_NV36ads_A10_v5 -Standard_NV6ads_A10_v5 Standard_NV72ads_A10_v5 2.5 2 1.5 1 0.5 0 s and and and the state of the start of the state of the state 1 2 8 2 8 2 all of the 如上图azure云 A10 价格最近还在上 租赁价格 如上图AWS A10租赁价格 还比较好。 ...

Company Developments - DeepSeek's R1 reasoning model research paper, co-authored by Liang Wenfeng, has been featured on the cover of the prestigious journal Nature, marking it as the first mainstream large language model to undergo peer review [2] - The global first AI Agent marketplace, MuleRun, developed by Alibaba's team, has officially launched, providing a platform for AI digital labor [2] - Magna International has signed a vehicle assembly contract with Xiaopeng Motors for the European market, marking Magna's first assembly project for a Chinese automaker, with production set to start in Q3 2025 [2] Market Expansion - Geely's Galaxy Starship 7 EM-i has officially launched in Australia, marking the second smart electric vehicle from Geely in the Australian market, with a sales growth rate exceeding 50% [3] - Didi's subsidiary 99 announced a 2 billion Brazilian real (approximately 2.6 billion yuan) investment in its food delivery platform 99Food, aiming to expand its services to 15 cities by the end of the year [4] - Keeta, Meituan's international food delivery brand, has launched operations in Kuwait, following its success in Saudi Arabia and Qatar [4] Partnerships and Collaborations - Grab has partnered with WeRide to launch autonomous driving services in Singapore, with an initial fleet of 11 vehicles [3] - WeRide and Pony.ai have announced plans to introduce fixed-route autonomous driving services in Singapore, pending regulatory approval [3] - The Saudi Central Bank has signed an agreement with Ant Group to launch Alipay+ cross-border payment services in Saudi Arabia by 2026 [5] Financing Activities - Yilujigou has completed a Series B financing round, raising several million yuan to expand its overseas warehouse network [6] - Enruikainuo has completed over 200 million yuan in Series A financing to accelerate innovative drug development and global expansion [6] - Qingyun New Materials has completed a Series C financing round, focusing on the development of new super materials and global capacity expansion [7] Regulatory Developments - Thailand's Trade Competition Commission is advancing new regulatory guidelines for digital e-commerce platforms, aiming to prevent market abuse and ensure fair competition [8]

机器之心报道 机器之心编辑部 大厂之间不是「你挖我」，就是「我挖你」。 那边特斯拉 Optimus AI 团队负责人 Ashish Kumar 被挖去 Meta，这边谷歌 DeepMind 资深研究科学家被 xAI 挖走了。 马斯克发推祝贺，并用火箭符号喊话：「起飞啦」！ 此次， 被挖去 xAI 的是一名在谷歌 DeepMind 工作近 9 年的大神级人物 ——Dustin Tran，离职前担任资深首席研究员 。 他是谷歌 Gemini-0801 的共同创造者，这是谷歌首个在 LMSYS 上登顶的模型。同时是 Gemini 2.5 系列模型的评测专家，这些模型在 WebDev Arena 和 HLE 等榜单 上取得了第一名。他还是谷歌 Gemini 1、1.5、2 和 2.5 的核心贡献者之一，其工作涵盖了强化学习、评测与数据等基础环节，并共同主导了相关论文与成果发布。 他在 X 上发表了一篇公开离职信，全文如下： 我在谷歌 DeepMind 工作 8 年多后选择了离开。这里留下了许多美好的回忆，最初在 Google Brain 参与早期奠基性的论文，与 Noam Shazeer、Ashish Vaswani ...

Core Viewpoint - The article discusses the intense competition for top AI talent among tech giants, highlighting significant financial incentives and strategic acquisitions that shape the AI landscape. It focuses on the case of Character.AI, which, despite losing its founders to Google, managed to achieve impressive revenue growth under new leadership while facing ongoing operational challenges and potential sale discussions [4][8][15]. Group 1: Talent Acquisition and Market Dynamics - Tech giants are increasingly willing to pay exorbitant sums for AI talent, exemplified by Google's $2.7 billion acquisition of Character.AI's founders and core team [10][12]. - The acquisition strategy often involves securing technology licenses to mitigate antitrust scrutiny while eliminating competition [10][11]. - The trend of "talent acquisition" reflects a harsh reality in the AI industry, where large companies systematically absorb promising startups and their talent, potentially stifling independent innovation [15]. Group 2: Character.AI's Transition and Performance - Following the departure of its founders, Character.AI was taken over by approximately 70 employees who demonstrated resilience and strategic focus, leading to a significant increase in monthly active users to over 20 million [17][18]. - The company shifted its strategy to focus on consumer products, leveraging open-source models to reduce operational costs while still aiming for profitability through subscription services [18][19]. - Character.AI's projected annual revenue is expected to reach $50 million by the end of 2025, up from a previous estimate of $30 million [18]. Group 3: Ongoing Challenges and Future Prospects - Despite its recent successes, Character.AI faces high operational costs, estimated in the millions per month, and regulatory pressures from lawsuits and investigations regarding harmful content [21][22]. - The company is exploring options for either a sale or new funding to sustain operations and improve its product offerings, with discussions about raising several hundred million dollars at a valuation exceeding $1 billion [22].