AI Daily: DeepSeek Image Recognition Mode Beta Test; Xiaohongshu Establishes AI Primary Department; Alibaba Launches Programmer Digital Avatar QoderWake

Welcome to the "AI Daily" section! This is your guide to exploring the world of artificial intelligence every day. Every day, we present you with the latest content in the AI field, focusing on developers to help you understand technology trends and innovative AI product applications.

Fresh AI products Click to learn more:https://app.aibase.com/zh

1. DeepSeek Launches Gray-scale Testing for Image Recognition Mode, Marking the Practical Application of Multimodal Visual Understanding

After releasing DeepSeek-V4, DeepSeek quickly initiated a gray-scale test for its multimodal image recognition function, marking a substantive phase in its multimodal capabilities. This feature added an "Image Recognition Mode" entry on mobile and web versions, showing strong performance in basic visual understanding, scene description, and logical reasoning, but still has room for improvement.

AiBase Summary:

✨ DeepSeek initiates gray-scale testing for multimodal image recognition, promoting the development of visual interaction.

🔍 The model performs excellently in basic visual understanding and logical reasoning, with high accuracy.

⚠️ Its recognition rate is limited when facing extreme visual challenges, requiring further optimization to improve accuracy.

2. Wenxin Yanyi 5.1 Preview Version Launched on LMSYS Arena, Ranking 13th Globally

The preview version of Wenxin Yanyi 5.1 was quietly launched on the international large model blind testing platform LMSYS Chatbot Arena, ranking 13th globally. This marks that Baidu's core model has entered a new cycle of rapid iteration and begun to accept direct quality inspection and benchmarking from global users.

AiBase Summary:

🚀 The preview version of Wenxin Yanyi 5.1 was quietly launched on the international large model blind testing platform LMSYS Chatbot Arena.

📊 This version ranks 13th overall, showcasing strong technical capabilities.

🌐 Baidu verifies its practical capabilities through public international mainstream evaluation systems, accelerating globalization efforts.

3. Xiaohongshu Establishes AI Primary Department "Dots"

Xiaohongshu announced the establishment of the AI primary department "Dots" and the Enterprise Intelligence Department, to enhance investment in artificial intelligence technology, and set up overseas departments "Rednote" and the Lab 1327 team, to promote international business and new product incubation.

AiBase Summary:

✅ Established AI primary department "Dots", covering multiple aspects including AI model R&D, infrastructure, engineering implementation, and product application.

✅ The Enterprise Intelligence Department integrates the original enterprise efficiency department and data science department, providing organizational structure and talent support for AI era development.

✅ Established overseas departments "Rednote" and Lab 1327 team, to promote international business and new product incubation.

4. Programmers Get Digital Twins: Alibaba Launches QoderWake, Achieving Fully Automated Code Repair Process

Alibaba launched QoderWake and Qoder mobile app, marking full scenario coverage of its AI agent ecosystem, enhancing R&D and operations automation level, and promoting AI to transition to a native operating system level.

AiBase Summary:

💻 QoderWake as a production-level digital employee, can independently execute tasks such as code change brief summary, error diagnosis, and generate repair code.

📱 Qoder mobile app supports cross-end collaboration and interactive experience innovation, users can remotely control the desktop end Agent to perform complex tasks via their phone.

🔄 Alibaba's Qoder product layout promotes AI from being a supporting tool to becoming a production factor with independent task processing capability.

Details link: https://qoder.com/qoderwake

5. Ant Group Officially Opens-Source Trillion-Level Large Model Ling-2.6-1T

Ant Group officially opens-source trillion-level large model Ling-2.6-1T, which optimizes instruction execution, tool adaptation, and long context handling through an innovative hybrid architecture, improving intelligence efficiency. It can also adapt to complex business scenarios involving multiple tools and constraints, demonstrating strong multi-step execution capabilities. In code generation, defect repair, and accurate reasoning under noisy environments, it has reached top levels in the open-source field.

AiBase Summary:

🧠 Ling-2.6-1T uses a hybrid architecture to improve intelligence efficiency.

🛠️ Supports complex business scenarios with multiple tools and constraints.

🚀 Reaches top levels in code generation and accurate reasoning in the open-source field.

Details link: https://huggingface.co/inclusionAI/Ling-2.6-1T

6. Juru Lu Announces Deep Collaboration with Volcano Engine, AI Short Plays Enter the "Industrialization" Era

Hangzhou Juru Lu Technology Co., Ltd. has reached a deep collaboration with Volcano Engine, integrating the Doubao video generation model Seedance 2.0, marking the entry of AI drama production into the industrialization era. By leveraging Volcano Engine's computing power and algorithm advantages, Juru Lu has achieved dual breakthroughs in production efficiency and image usability, and built a full-stack technical architecture, driving the domestic AI film industry toward a more mature industrialized stage.

AiBase Summary:

🚀 Double leap in efficiency and quality: AI drama production efficiency increases nearly ten times, and the production cycle is compressed from 15-30 days to 1-3 days.

🖼️ Image usability has significantly improved: the pass rate of images in traditional AI generation mode was only 30%, while the new technical architecture has increased it to over 90%.

🛠️ Full-stack technical architecture: The two parties have jointly built a technical system covering pre-production creation to final delivery, lowering the AI drama production threshold and ensuring high-quality content.

7. Say Goodbye to Copy-Paste! Gemini Receives a Historic Update, One-Click Generation of Office Documents

Gemini has significantly enhanced its capabilities as a productivity tool by adding the ability to directly generate and export files in various formats, while also improving compatibility with office software, providing users with a more efficient work experience.

AiBase Summary:

✨Gemini adds the ability to directly generate and export files in various formats, enhancing work efficiency.

📊 Supports mainstream document formats such as Google Docs, Word, Excel, with strong compatibility.

🖼️ Introduces image recognition features, capable of converting handwritten notes into well-formatted PDF files.