AI Daily: Shengshu Technology Releases Vidu Q4 Preview Version; Alibaba Qoder Launches Fast Mode for the First Time; Tencent Cloud Open Sources TeamAI

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 technical trends and innovative AI product applications.

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

1. Shengshu Technology launches Vidu Q4 Preview with 4K output support

Shengshu Technology released Vidu Q4 Preview, significantly improving video generation quality, supporting multi-modal references and high-resolution output, providing creators with more tools and higher quality video content.

【AiBase Summary:】

🎥 Supports image-to-video and reference-to-video modes, enhancing creative flexibility.

🔊 Multi-modal reference capacity breakthrough, supporting 3 audio clips and 15 images, ensuring consistency in visuals.

📺 Optimized dynamic camera movement and complex shot performance, supports 4K resolution output, meeting professional film and television needs.

2. Alibaba Qoder undergoes major upgrades: Launching Fast mode, response time only 3 seconds and reduced costs

Alibaba Qoder undergoes major upgrades: Launching Fast mode, response time only 3 seconds and reduced costs.

【AiBase Summary:】

🚀 Alibaba Qoder launched Fast mode, improving response speed and efficiency.

💡 CN version adds a Fast tier, international version upgraded to Fast mode.

📉 Credit consumption reduced by 27%, demonstrating competitiveness.

3. Tencent Cloud open-sources TeamAI: Compatible with 16 Agents, team Skills go through Git review and distribution, cost reduced by 76%

Tencent Cloud open-sourced the AI collaboration tool TeamAI, which has been used internally and appeared on GitHub Trending. TeamAI manages team skills, work standards, and tool configurations based on Git, allowing one person to configure and everyone to reuse, and it is compatible with multiple Agents. Testing showed that models with access to team knowledge improved by 6.1% and call costs were reduced by 76%.

【AiBase Summary:】

✨ TeamAI uses Git for unified management of team skills and tool configurations, supporting 16 Agents.

🔧 Through Git review and distribution mechanisms, achieve efficient reuse and version control of configurations.

📈 Testing shows that TeamAI can significantly improve model scores and reduce call costs.

4. AI reply content appears in printed novels, publisher responds: will recall and destroy and re-edit

The printed novel "The Last Road of the End" unexpectedly featured AI reply content, sparking online discussion. The publisher responded that the issue originated from an oversight in the editing process, and has decided to recall and destroy the affected books and re-edit them for publication.

【AiBase Summary:】

📘 Printed novel unexpectedly features AI reply content, drawing attention.

🔍 Publisher responded that the issue stemmed from an oversight in the editing process, not the author's responsibility.

🗑️ Publisher has decided to recall and destroy the affected books and re-edit them for publication.

5. Google Gemini 4 is about to be released, mysterious new version aims for the strongest code model

Google is about to release the new Gemini4 model Argon, and is testing its subsequent iteration Carbon. Carbon shows strong performance in code writing and processing, even being compared to Anthropic's Opus5.5. However, it remains unclear whether Carbon will be released as a fine-tuned version of Argon or as an independent model.

【AiBase Summary:】

🔥 Google is about to release the new Gemini4 model Argon.

💡 Internal tests show that Carbon performs well in code processing, possibly matching Opus5.5.

❓ Whether Carbon is a follow-up version of Argon or an independent model is still unclear.

6. Tencent Yuanbao introduces desktop pet: PC version v2.87.10 launched, offering round-the-clock companionship + drag-and-drop file summary

Tencent Yuanbao introduced a new desktop pet function on the PC, enhancing user experience with round-the-clock companionship and smart task reminders. This feature supports drag-and-drop file, image, and text for quick processing, and has the ability to stick to the edge and initiate conversation, helping users work more efficiently.

【AiBase Summary:】

⏰ Tencent Yuanbao's desktop pet is now available on PC, supporting round-the-clock companionship and task reminders.

📁 Supports drag-and-drop files, images, and text, enabling quick summarization and processing.

🧠 Desktop pets have the ability to stick to the edge and initiate conversation, enhancing user interaction experience.

7. Youdao Longxia releases migration assistant: Competing for Tencent QClaw users, offering one-click "migration"

Youdao Longxia issued an open letter regarding the shutdown of Tencent QClaw, emphasizing its strategic positioning as a desktop Agent product for office users and inviting QClaw users to "migrate." Since its launch in February 2026, Youdao Longxia has iterated 82 versions, adhering to open-source, open, and security principles, emphasizing long-term investment and user data value.

【AiBase Summary:】

📌 Youdao Longxia responds to the shutdown of Tencent QClaw, clarifying its strategic positioning.

📌 Offers a convenient migration solution, capturing market gaps and strengthening its position in the AI office assistant sector.

📌 Emphasizes long-term investment and user data value, not using user experience as a internal competition tool.

8. Claude Management Agent launches dynamic workflow: Can schedule up to 1000 AI Agents at once

Anthropic announced a new dynamic workflow feature for the Claude Management Agent, significantly enhancing the ability of AI agents to handle complex tasks, with the capability to schedule up to 1000 agents simultaneously, suitable for scenarios such as large codebase checks and massive data cleaning. Testing showed that dynamic workflows outperformed traditional single agents in code vulnerability detection, effectively improving efficiency and stability.

【AiBase Summary:】

✅ Dynamic workflow feature enhances the parallel processing capabilities of AI agents, capable of coordinating up to 1000 agents.

🔍 Dynamic workflows were validated for efficiency and stability through code vulnerability detection testing.

🚀 Marking a shift from "solo operations" to "thousand-person collaboration" in enterprise-level tasks for AI agents.