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 learn about innovative AI product applications.

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

1. WeChat Reading Integrates AI Skill, Helping You Organize Your Reading Life in One Sentence

The article introduces the new AI Skill feature of WeChat Reading, which allows the intelligent assistant to help users manage their bookshelves, analyze reading habits, and recommend books by authorizing accounts, thus improving reading efficiency and experience.

image.png

AiBase Summary:

📚 Manage Your Bookshelf: View all your books at a glance, avoiding the hassle of not finding a book.

📊 Reading Statistics: Automatically analyze reading duration, days, and preferences, making habits visual.

✨ Recommend Great Books: Push booklists that match your real preferences.

More Details: https://weread.qq.com/r/weread-skills

2. Monthly AI Computing Power Package! Major Telecommunications Operators Join the Token Plan Launch

The article introduces the Token plans launched by major telecom operators in the era of large AI models, offering AI computing power in a way similar to mobile data packages, breaking through the previous barriers where only internet giants had exclusive model billing.

image.png

AiBase Summary:

💰 Monthly AI Computing Power Package! All major telecom operators have joined the Token plan launch.

📱 Token plans support direct payment with phone credit and cross-platform compatibility.

🌐 Major telecom operators are accelerating their transformation toward computing power.

3. Alibaba Tongyi Qianwen Makes a Major Upgrade: New Flagship Qwen3.7 Preview Version Debuts Globally

The large model of Tongyi Qianwen under Alibaba Cloud has undergone a key iteration. The newly released Qwen3.7-Max-Preview and Qwen3.7-Plus-Preview have shown excellent performance in logical reasoning and visual fields, demonstrating its leading position in the global technology field.

image.png

AiBase Summary:

🧠 The new models Qwen3.7-Max-Preview and Qwen3.7-Plus-Preview have been launched, focusing on high-level logical reasoning and deep computation.

📊 In the text processing field, Qwen3.7 Max Preview ranks 13th globally, helping Tongyi Qianwen enter the top six globally.

🖼️ Qwen3.7 Plus Preview ranks 16th globally in the visual field, and the lab's overall ranking jumps to fifth globally.

4. No More Parameter Wars, But Real-world Applications! Sohu Clearly Embraces the "Rational Application" AI Strategy: How Can Mid-sized Tech Companies Find a Differentiation Path Between Giants?

Sohu chooses not to participate in the arms race of large models but instead focuses on applying large models to actual business, improving efficiency and maintaining content neutrality. This strategy reflects the thinking of mid-sized tech companies to achieve differentiation through precise positioning under limited resources.

image.png

AiBase Summary:

✅ Improve Efficiency and Reduce Costs: Use mature large models from top vendors to empower core processes such as content production, review, and distribution, making the process lighter and increasing human efficiency.

✅ Maintain Content Neutrality: Uphold an objective stance in algorithm recommendations and intelligent generation to avoid over-迎合ing traffic and maintain the credibility of media platforms.

✅ Focus on Vertical Scenarios: Transform general capabilities into business value, such as semantic understanding of news information, intelligent editing of video content, and personalized responses to user interactions.

5. Alexa+ Launches On-demand Podcast Function - Just Tell It What You Want!

Alexa+ launches an on-demand podcast function, allowing users to quickly generate personalized podcasts with simple commands. The AI automatically handles content creation and adjustment and collaborates with authoritative news agencies to improve content quality.

image.png

AiBase Summary:

🎙️ Alexa+ new feature supports on-demand generation of podcast programs, allowing users to create personalized podcasts with just simple commands.

🤖 AI will automatically research topics, collect information, and generate content outlines, which users can adjust anytime regarding duration and content focus.

📰 Amazon collaborates with multiple authoritative news agencies to ensure the accuracy and credibility of generated content.

6. More Features, Same Price! Anthropic's Two Main AI Tools Receive Major Upgrades

Anthropic has upgraded its AI tools Claude Design and Claude Code, improving users' creative efficiency and development experience.

image.png

AiBase Summary:

🧠 The token usage limit of Claude Design has doubled, solving the pain points for creators handling complex design projects.

💻 Claude Code defaults to the more advanced Opus 4.7 model, with response speed increased by 2.5 times.

💰 The pricing of Opus 4.7 in fast mode remains the same as the previous version, offering ultra-long context window support.

7. Robot Interaction No Longer "Naked"! Zhiyuan WITA Completes Compliance Filing First

The article introduces WITA, a large model from Zhiyuan, which has completed filing, becoming the first compliant落地 body-intelligent interaction large model in the country. The core mission of WITA is to make humanoid robots truly "able to chat, understand emotions, and have personalities," and it has already been applied in commercial scenarios such as tour guiding and retail services. In addition, Zhiyuan also launched the "358 Vision Plan," aiming to exceed 10 billion yuan in revenue by 2027 and 100 billion yuan by 2030.

image.png

AiBase Summary:

🤖 WITA large model has completed filing, becoming the first compliant body-intelligent interaction large model in the country.

💬 WITA Omni1.0 will realize end-to-end multi-modal interaction, enhancing robot interaction experience.

📈 Zhiyuan launched the "358 Vision Plan," aiming to exceed 10 billion yuan in revenue by 2027.

8. Cursor Releases Composer2.5 Coding Model: Competing with GPT-5.5 and Opus4.7 at a Very Low Cost

Cursor released the Composer2.5 coding model, whose performance is comparable to GPT-5.5 and Opus4.7, while significantly reducing costs, showing a high cost-performance ratio and commercial competitiveness.

image.png

AiBase Summary:

🧠 Composer2.5 is built based on Moonshot's Kimi K2.5 checkpoint, with a training task scale 25 times larger than the previous version.

💰 Composer2.5 costs less than $1, far lower than the competitors' $11, showing strong commercial competitiveness.

🚀 Cursor is collaborating with SpaceX and xAI to advance the training of even larger subsequent models from scratch.