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.

New AI Products Learn More: https://top.aibase.com/

1. Zhipu Releases Agent Product AutoGLM Contemplation: The First Intelligent Agent to Think and Act Simultaneously

Zhipu AI released its latest Agent product, AutoGLM Contemplation, at the 2025 Zhongguancun Forum. This product is the industry's first intelligent agent that integrates in-depth research and operational capabilities, capable of performing complex thinking and executing actions simultaneously. AutoGLM Contemplation is based on Zhipu's self-developed GLM full-stack large model, marking a step towards more practical and user-friendly intelligent agent technology. It is expected to bring significant breakthroughs in research analysis and information processing.

image.png

【AiBase Summary:】

🧠 AutoGLM Contemplation is the first intelligent agent capable of thinking and acting simultaneously, handling complex thought processes and actions concurrently.

🌐 This product can browse web pages, retrieve data, analyze information, and generate reports, much like a human.

🚀 Zhipu's self-developed GLM full-stack large model provides powerful technical support for AutoGLM Contemplation, resulting in significant performance and speed improvements.

2. Google Releases New AI Model Gemini 2.5 Pro for Free Use

Google's latest Gemini 2.5 Pro AI model is now available for free to all users. Called the most intelligent AI to date, it boasts features like app and browser extensions, file uploads, and seamless integration with Google's collaborative tool, Canvas. Although still in the experimental phase, Gemini 2.5 Pro's powerful reasoning capabilities and 1 million token context window make it excel at handling large datasets and extended conversations. Google plans to further expand its capabilities.

image.png

【AiBase Summary:】

🆓 Gemini 2.5 Pro is now free for all users, removing previous paywalls.

📈 The new model shows significant improvements in reasoning capabilities, supports various functions, and allows file uploads.

🔗 Gemini 2.5 Pro has a 1 million token context window, with future plans to expand to 2 million tokens.

3. ChatGPT's Native Image Generation Feature Quietly Released to Free Users

OpenAI's recently released ChatGPT image generation feature has garnered significant attention, particularly its "Ghibli-style" image generation. Previously only available to paying users, this feature is now accessible to all free users, allowing them to transform ideas into visual creations. However, its popularity has sparked controversies surrounding copyright and plagiarism, especially concerning works by Studio Ghibli.

image.png

【AiBase Summary:】

🌟 ChatGPT's image generation feature is now available to all free users, offering a novel creative experience.

⚖️ The release has sparked debates about copyright and plagiarism, particularly concerning Studio Ghibli's works.

🔧 OpenAI states that due to high usage, temporary speed limits will be implemented to address GPU overheating issues, with plans to grant free users three generation attempts per day.

4. Say Goodbye to Node Nightmares! ComfyUI-C opilot Released, Featuring GPT-4-like Image Generation and Editing Capabilities

ComfyUI-C opilot is an innovative AI tool that combines natural language processing with node-based workflows, significantly lowering the barrier to entry for image generation. Users simply describe their needs, and the tool automatically builds the workflow, providing an efficient creative experience. Supporting both Chinese and English interaction, it also provides real-time resource recommendations and problem-solving, making it a valuable assistant for designers and content creators.

image.png

【AiBase Summary:】

🌟 ComfyUI-C opilot simplifies complex AI image generation processes through natural language interaction, allowing users to easily get started without technical expertise.

🤖 The tool integrates several practical functions, capable of recommending models in real-time, explaining node functions, and providing repair suggestions when problems arise.

🚀 The launch of ComfyUI-C opilot is considered a revolution in AI image generation tools, significantly improving creative efficiency and boosting the productivity of designers and content creators.

5. Fliggy Unveils AI Achievements: Online Customer Service Efficiency Increases by 54%, Merchant Productivity Doubles

Fliggy recently revealed significant progress in artificial intelligence. AI-assisted coding penetration has reached 70%, and all hotel business personnel are equipped with intelligent agents, significantly improving business development and customer service efficiency. AI customer service has handled approximately one-tenth of user inquiries, increasing response efficiency by 54%. Additionally, Fliggy has launched a series of AI operational tools to help merchants achieve partial automation, improving overall operational efficiency.

image.png

【AiBase Summary:】

🌟 AI-assisted coding penetration rate reaches 70%, with 100% of hotel business personnel equipped with intelligent agents.

🤖 Online customer service efficiency increased by 54%, with AI customer service handling about one-tenth of user inquiries.

🚀 Fliggy plans to use AI as a driver over the next three years to achieve a breakthrough in both the quality and scale of its supply chain.

6. Wayve Launches Comprehensive World Model GAIA-2: Supports Simultaneous Generation of Five-View Videos, Can Simulate High-Risk Scenarios

Wayve has launched the GAIA-2 video generation world model, significantly improving the training and validation capabilities of autonomous driving systems. This model can generate diverse driving scenarios and supports the simultaneous generation of five-view videos, enhancing environmental perception. Furthermore, GAIA-2 can simulate high-risk scenarios, helping developers verify the safety of autonomous driving systems in a controlled environment.

image.png

【AiBase Summary:】

🌍 GAIA-2 enhances AI model adaptability to different traffic rules by expanding geographical coverage and generating driving scenarios from multiple countries.

🎥 Supports the simultaneous generation of up to five video perspectives, ensuring consistency across multiple camera views and enhancing environmental perception.

⚠️ Capable of simulating high-risk scenarios, helping developers validate the robustness and safety of autonomous driving systems in a controlled environment.

Details: https://top.aibase.com/tool/gaia-2

7. AI-Powered Paper Search Ai2 PaperFinder: Input Topic, AI Automatically Searches All Relevant Papers

In modern scientific research, Ai2PaperFinder, a powerful literature retrieval tool, aims to help researchers efficiently access relevant academic papers. This platform contains over 8 million full-text papers and 108 million abstracts. Users can quickly find the literature they need through simple keyword searches. Ai2PaperFinder also features literature synthesis capabilities, providing users with detailed answers and significantly saving research time.

image.png

【AiBase Summary:】

🌟 Ai2PaperFinder contains over 8 million full-text papers and 108 million abstracts, providing powerful literature retrieval capabilities.

📚 Users can quickly find relevant research through simple keyword searches, whether for popular or niche topics.

🛠️ The platform is developing new tools to further enhance user experience and help researchers better integrate and share literature.

Details: https://top.aibase.com/tool/ai2-paperfinder

8. Goodbye to Stiff AI! Midjourney and NYU Unlock New Dimensions in Creative Text, Diversity Soars by 23%!

Midjourney and researchers from New York University have developed a new method that significantly improves the diversity of creative text generated by language models while controlling quality loss. This technology introduces a "bias metric" to quantify the differences in generated text, enhancing output diversity. Initial tests show a 23% increase in model text diversity with only a 5% decrease in quality score.

image.png

【AiBase Summary:】

🌟 The new method, using a bias metric, enhances the diversity of creative text generated by language models.

📈 Initial tests show a 23% increase in text diversity with only a 5% quality loss.

🛠️ The code will be publicly released for other researchers and developers.

9. Alibaba Cloud and Nanjing University Jointly Launch AI Talent Cultivation Program

Nanjing University and Alibaba Cloud's collaborative program aims to cultivate future AI innovators, emphasizing cross-disciplinary thinking. Through the design of forward-looking AI courses and practical training, both parties will promote the deep integration of interdisciplinary studies and talent cultivation. Alibaba Cloud also provides Nanjing University students with free cloud computing and AI resources to support their teaching and research activities.

image.png

【AiBase Summary:】

🌟 Nanjing University and Alibaba Cloud launch a joint AI talent cultivation program aimed at fostering cross-disciplinary AI innovators.

💻 Alibaba Cloud provides Nanjing University students with free cloud computing and AI resources to support teaching and research.

🤝 Both parties will jointly conduct academic exchanges, promote technological achievements transformation, and open up new models for new engineering talent cultivation.

10. Tencent’s HunYuan Large Model Application Practical Course Officially Launched on the National Smart Education Platform

Tencent's recently launched "Tencent HunYuan Large Model Application Practical Course" is now available on the National Smart Education Platform, aiming to provide college students and teachers with comprehensive knowledge and practical skills in AI large model technology. This course complements the Ministry of Education's Smart Education Platform 2.0 intelligent version, promoting the digital transformation of education. The course content covers theory and practice, promoting interaction between teachers and students, stimulating creativity, and contributing to the innovative development of future education and research.

image.png

【AiBase Summary:】

🎓 The course provides comprehensive knowledge of AI large model technology, helping teachers and students efficiently apply it to education and research.

🤝 Promotes interaction and collaboration among college teachers and students, stimulating creativity and research enthusiasm.

🚀 Provides impetus for the digital transformation of education, helping teachers and students master emerging skills and meet future challenges.

11. Embodied AI Company Qixun Intelligent Completes 528 Million Yuan Pre-A Round of Financing

Qixun Intelligent has made significant progress in embodied AI financing, successfully completing a Pre-A round of financing of 528 million RMB, demonstrating the high recognition of its technological strength and market prospects by the capital market. The company is committed to promoting the development of embodied AI technology, particularly innovation in the field of robotics, marking its growing competitiveness in the global market.

image.png

【AiBase Summary:】

💰 This round of financing was led by Prosperity7 Ventures, under Aramco Ventures, with participation from several well-known capital firms including China Merchants Venture Capital.

🤖 Qixun Intelligent is committed to building world-class general-purpose humanoid robots and has successfully developed the Moz1, a commercial-grade humanoid robot with high-precision full-body force control.

🚀 The company has achieved key technological breakthroughs in the AI field, successfully realizing long-range manipulation of flexible objects, marking significant technological progress.