Welcome to the "AI Daily" column! Here is your guide to exploring the world of artificial intelligence every day. Every day, we bring you the latest hot topics in the AI field, focusing on developers, helping you understand technical trends and learn about innovative AI product applications.
Fresh AI products Click to learn more:https://app.aibase.com/zh
1. Meituan LongCat-Flash-Omni officially released, opening a new era of full-modal real-time interaction
Meituan's LongCat-Flash-Omni model has made significant breakthroughs in full-modal real-time interaction. It uses the latest ScMoE technology and performs well in multiple fields, providing developers with efficient multimodal application scenario solutions.
AiBase Highlights:
🧠 Integrates an efficient multimodal perception module and voice reconstruction module
🚀 Uses Shortcut-Connected MoE technology to achieve low-latency real-time audio-visual interaction capabilities
🌐 Supports full-modal tasks, showing excellent performance in text, image, video understanding, and speech perception and generation
Details link: https://huggingface.co/meituan-longcat/LongCat-Flash-Omni
2. Alibaba Tongyi Qwen3-Max launches deep thinking function on its official website
Alibaba's latest flagship language model, Qwen3-Max, officially launched the 'Deep Thinking' mode, significantly improving the efficiency of handling complex tasks. The model's parameter count exceeds 1 trillion, with pre-training data reaching 36T tokens, and it shows excellent performance in multiple benchmark tests, demonstrating strong reasoning and programming capabilities.
AiBase Highlights:
🧠 Qwen3-Max is Alibaba Tongyi's latest flagship language model, with parameters exceeding 1 trillion.
🔍 The newly launched 'Deep Thinking' mode enhances the ability to analyze reasoning chains and decompose multi-step problems.
🏆 Qwen3-Max-Thinking version achieved 100% accuracy in high-difficulty reasoning benchmark tests.
3. Baidu "Wenxin" 5.0 makes a big comeback! One-click creation of comics, photo editing, videos, all-in-one AI assistant upgraded comprehensively
The article details the many functional upgrades of Baidu's AI assistant "Wenxin" 5.0, including magical comics, creative photo editing, "Safe Writing", full-modal interaction, video generation, and multilingual communication, showcasing its powerful capabilities as an all-in-one AI platform.
AiBase Highlights:
🎨 Magical Comics: Users upload photos and input descriptions to generate coherent comics
🖼️ Creative Photo Editing: Intelligent photo editing engine supports artistic filters and style transfer
🎥 Video Generation: Static images can be converted into dynamic videos and support multilingual communication
4. Cloud storage acceleration: Baidu Netdisk core API compatible with MCP protocol, empowering developers to access with one click
Baidu Netdisk upgraded its core API by compatibility with the MCP protocol, significantly simplifying the developer access process and enhancing file management and retrieval capabilities, injecting new vitality into the cloud storage industry.
AiBase Highlights:
📎 Baidu Netdisk's core API is fully compatible with the MCP protocol, simplifying the developer access process.
🔍 Provides efficient file search functions, supporting semantic search and various file operations.
🔄 Enhances upload methods to meet data access needs in different scenarios.
Details link: https://github.com/baidu-netdisk/mcp
5. OpenAI opens Sora2 video tool, available to users in the US, Canada, Japan, and South Korea
OpenAI announced the removal of Sora2's invitation code restrictions, officially making it available for download to users in the US, Canada, Japan, and South Korea, marking its first large-scale expansion and entry into the Asian market. At the same time, to address resource shortages, it introduced a $4 "credit pack" to increase generation quotas and plans to build a "Sora economy," charging per use for appearances of copyrighted characters and famous people, responding to controversies about "default collection."
AiBase Highlights:
🌍 OpenAI opens Sora2 video tool, available to users in the US, Canada, Japan, and South Korea.
💰 Introduces a $4 "credit pack" to accelerate commercialization and provide additional generation capacity.
📜 Plans to build a "Sora economy," planning to charge for appearances of copyrighted characters and famous people.
6. Google CEO confirms: Gemini will be released within 3 years, AI Agent capabilities may become the breakthrough point
Google CEO Sundar Pichai confirmed during the earnings call that the company plans to launch the next-generation AI model, Gemini3, within the year. This model will focus on enhancing the 'agent' capabilities for handling complex, multimodal tasks, aiming to close the gap with competitors like OpenAI's GPT-5. Meanwhile, Alphabet's quarterly revenue exceeded $100 billion for the first time, showing the significant role of AI technology in business growth.
AiBase Highlights:
🚀 Gemini3 focuses on enhancing multimodal task and agent capabilities to improve performance.
💰 Alphabet's quarterly revenue exceeded $100 billion for the first time, with AI becoming a core growth driver.
🤝 Deepening cooperation: Anthropic plans to use 1 million Google TPUs for model training, showing the appeal of Google's AI infrastructure.
7. Siri is about to make a comeback? Apple will release a major update "Apple Intelligence" in March next year, using Google's Gemini!
Apple plans to launch a new generation of Siri in March 2026, introducing Google's Gemini large model technology, and pairing it with a new smart home display device. It will also fully showcase the Apple Intelligence strategy at WWDC, achieving a smart leap.
AiBase Highlights:
🍎 Introduces Google's Gemini large model technology to enhance Siri's web understanding and real-time information retrieval capabilities.
🏠 Launches a new smart home display device, becoming the core entry point for family AI interaction.
📅 In 2026, WWDC will fully integrate Apple Intelligence capabilities, building an end-to-end personal intelligent ecosystem.
8. Generate AI Agents in one sentence! Pokee AI's no-code solution sparks an automation revolution, threatening OpenAI and n8n?
Pokee AI enables no-code AI Agent development through natural language instructions, greatly simplifying traditional complex processes and driving an automation revolution.
AiBase Highlights:
🤖 Create intelligent workflows through natural language instructions without any programming skills.
🧠 Self-developed "prompt to workflow" engine supports interactive logic preview and adjustment.
🌐 Compatible with thousands of mainstream applications, enabling cross-platform automation operations.
