Welcome to the 【AI Daily】column! This is your daily guide to exploring the world of artificial intelligence. Every day, we present you with the hottest AI news, focusing on developers and helping you understand technology trends and innovative AI product applications.
Discover new AI products Learn More: https://top.aibase.com/
1. Anthropic Launches Hybrid Reasoning Model Claude 3.7 Sonnet: Outperforming DeepSeek
Today, AI company Anthropic announced its latest hybrid reasoning model, Claude 3.7 Sonnet, claiming it to be the most intelligent AI model ever. Combining reasoning and traditional methods, the model efficiently handles complex problems while maintaining fast response times. Anthropic also released Claude Code, a programming tool, further expanding the applications of the Claude series.
【AiBase Summary:】
🧠 Claude 3.7 Sonnet is the only hybrid reasoning model on the market, combining reasoning and real-time generation capabilities.
💻 The newly launched Claude Code tool allows for code searching, editing, and GitHub commits, enhancing developer collaboration.
🌐 Users can access Claude 3.7 Sonnet through the Claude app, API, and other methods, at the same cost as previous versions.
2. Alibaba's Tongyi Qianwen Launches Inference Model QwQ-Max Preview Version, Available at qwen.ai
Alibaba recently launched QwQ-Max-Preview, an inference model based on Qwen 2.5-Max, and plans to fully open-source its latest inference model. This move will not only promote the popularization of AI technology but also provide developers and enterprises with more powerful reasoning capabilities, especially in code generation and complex task processing. The performance of QwQ-Max-Preview is comparable to OpenAI's o1-medium, demonstrating Alibaba's competitiveness in the AI field.
【AiBase Summary:】
🚀 QwQ-Max-Preview is Alibaba's latest inference model, planned for full open-sourcing.
📈 Performance evaluations show that QwQ-Max-Preview excels in inference speed and code generation accuracy.
🌐 Users can access and experience the latest inference model directly via the qwen.ai domain.
3. DeepSeek Open Source Week Day Two: The First Open-Source EP Communication Library for MoE Models
On the second day of its open-source week, DeepSeek launched DeepEP, a highly efficient communication library specifically designed for Mixture-of-Experts (MoE) models. DeepEP aims for full-stack optimization, supporting low-latency and high-throughput many-to-many GPU kernels, especially suitable for training and inference tasks. The library is not only compatible with low-precision operations like FP8 but also optimizes the bandwidth performance of data forwarding.
【AiBase Summary:】
🌟 DeepEP is designed for Mixture-of-Experts models, providing high-throughput and low-latency communication solutions.
⚙️ Supports various low-precision operations and optimizes data forwarding bandwidth performance.
💡 Tested and verified, DeepEP is compatible with InfiniBand networks and is suitable for isolation and management of different traffic types.
Details: https://x.com/deepseek_ai/status/1894211757604049133
4. Pinduoduo Forms E-commerce Recommendation Large Model Team: Leader's Salary Multiples Higher Than at Baidu, Internal "Horse Racing" Mechanism Implemented
Pinduoduo announced the formation of an e-commerce recommendation large model team, further promoting the application of AI technology in the e-commerce field. Led by a core member from Baidu's Phoenix Nest, the team boasts significantly increased salaries, attracting numerous AI experts. The team focuses on optimizing user experience and improving transaction efficiency, using a "horse racing" mechanism to incentivize technological iteration and aiming to challenge Alibaba and JD.com's market positions.
【AiBase Summary:】
💼 The leader's significantly increased salary reflects Pinduoduo's substantial investment in talent acquisition.
🔍 The team focuses on building a cutting-edge recommendation system covering price comparison, recommendations, advertising, search, and customer service.
🏆 The "horse racing" mechanism incentivizes team members to rapidly iterate technology and improve market competitiveness.
5. AI Programming Tool Cursor Integrates Claude 3.7 Sonnet Reasoning Model
Cursor AI recently announced the integration of the Claude 3.7 Sonnet reasoning model and a user interface update for improved user experience. Claude 3.7 Sonnet is widely recognized for its innovative design and programming capabilities. Users can choose between quick answers or expanded thinking modes, suitable for complex tasks and providing more accurate solutions.
【AiBase Summary:】
⚙️ Cursor AI integrates Claude 3.7 Sonnet, improving the user interface and experience.
💡 Claude 3.7 Sonnet combines quick answers and in-depth thinking, suitable for complex tasks.
📈 Sonnet demonstrated outstanding performance in programming tests, outperforming other well-known models.
6. ChatGPT Adds Safari Extension, Settable as Default Search Engine in Safari Browser Address Bar
OpenAI recently released a new version of the ChatGPT app, adding a Safari browser extension that allows users to set ChatGPT as their default search engine. This update aims to enhance the user search experience, making information retrieval more intelligent and personalized. Users can simply enable this function in the settings to use ChatGPT directly in the Safari address bar for searches.
【AiBase Summary:】
🔍 The new ChatGPT version allows users to set it as the default search engine for the Safari browser.
📲 Users can enable the ChatGPT search extension through settings for intelligent search.
💡 This feature enhances user experience and promotes the development of artificial intelligence in the search engine field.
7. SenseTime's Xiaohuanxiong Family Receives Comprehensive Upgrade: Multimodal Fusion
At the 2025 Global Developer Pioneer Conference, SenseTime announced a comprehensive upgrade to its AI productivity tool, the SenseTime Xiaohuanxiong family, strengthening its multimodal capabilities, accelerating AI application deployment, and evolving towards AI Agents. This upgrade not only improves tool performance but also enables AI to better serve productivity. The multimodal capabilities of Office Xiaohuanxiong and Code Xiaohuanxiong 2.0 demonstrate their strong potential in data analysis and programming efficiency, attracting a large number of user registrations.
【AiBase Summary:】
🚀 The upgraded SenseTime Xiaohuanxiong family has achieved a comprehensive improvement in multimodal capabilities, enabling efficient information integration and data analysis.
📈 Office Xiaohuanxiong, with automated tool support, easily handles complex tasks and improves team collaboration efficiency.
💻 Code Xiaohuanxiong 2.0 realizes multi-dimensional data fusion, enhancing programming efficiency and supporting real-time retrieval of technical documents.
Details: https://xiaohuanxiong.com/login
8. Anthropic's Next Funding Round May Increase to $3.5 Billion, Valuation Reaching $61.5 Billion
According to the Wall Street Journal, AI startup Anthropic is undergoing a new funding round, increasing from $2 billion to $3.5 billion, with a valuation reaching $61.5 billion. Despite an annualized revenue of approximately $1.2 billion, Anthropic remains unprofitable and plans to use this funding to increase investment in AI technology research and development. Several well-known investment firms are participating in this round of financing, which is expected to provide strong financial support for the company's future development.
【AiBase Summary:】
💰 Funding increased from $2 billion to $3.5 billion, with a valuation of $61.5 billion.
🤖 Annualized revenue is approximately $1.2 billion, but the company is still losing money and plans to increase investment in AI technology R&D.
🚀 Several well-known investment firms are participating in the funding, potentially providing stronger financial support for future development.
9. Baidu Education's "Photo Search for Questions" and "AI Essay Writing" Integrate DeepSeek-R1 Model
On February 25th, Baidu announced that its educational products "Photo Search for Questions" and "AI Essay Writing" officially integrated the DeepSeek-R1 model, marking a significant technological upgrade in the field of AI education. This integration not only improves the intelligence level of learning tools but also provides users with a more efficient learning experience, helping them master problem-solving methods and writing skills.
【AiBase Summary:】
📸 The "Photo Search for Questions" function integrates the DeepSeek-R1 model, providing standard answers and problem-solving processes to help users grasp problem-solving thinking.
✍️ AI essay writing achieves a shift from word piling to collaborative thinking, supporting personalized styles and secondary editing to inspire user creativity.
🔍 Baidu Search, after DeepSeek-R1 integration, significantly improves the search experience on both PC and mobile ends, assisting students in their learning progress.
10. JD Retail Technology Releases Jing Dian Dian AIGC Content Generation Platform: One-Click Generation of Product Images and Marketing Copy
The Jing Dian Dian AIGC content generation platform, launched by the JD Retail Technology team, marks a significant technological breakthrough in the field of e-commerce content generation. Using AI technology, the platform helps merchants efficiently and cost-effectively generate product images and marketing copy, covering over 20 core scenarios with a daily call volume exceeding 10 million.
【AiBase Summary:】
🖼️ The Jing Dian Dian AIGC platform is open, allowing merchants to generate high-quality product images and marketing copy with one click.
🤖 The platform uses AI technology to improve content production efficiency by over 95% and reduce costs by over 99%.
📈 Innovative technology ensures the realism and consistency of generated images, meeting diverse needs.
Details: https://ai.jd.com
