Welcome to the 【AI Daily】 section! This is your guide to exploring the world of artificial intelligence every day. Here, we present you with hot topics in the AI field daily, focusing on developers to help you gain insight into technical trends and understand innovative AI product applications.
Fresh AI products click to learn more: https://top.aibase.com/
1. $249.99 per month! Google launches AI Ultra subscription plan for access to the most powerful AI models and advanced features
Google has launched a new AI Ultra subscription plan, offering access to the highest level of AI models and rich advanced features, suitable for filmmakers, developers, and creative professionals. The plan includes the latest Gemini applications, Flow tools, 30TB of storage space, etc., with a subscription fee of $249.99 per month.
【AiBase Summary:】
🌟 Provides access to the highest-level AI models and rich advanced features, suitable for users seeking the best AI experience.
🎥 Flow tools support high-quality video generation, mastering advanced photography control and experiencing Veo3 functions.
💾 Offers 30TB of storage space, ensuring user creation and file security.
For more details: https://blog.google/products/google-one/google-ai-ultra/
2. Tencent Hunchun announces full upgrade of its model matrix, introducing new visual reasoning model T1-Vision and voice communication model Hunchun Voice
Tencent Hunchun has comprehensively upgraded its model matrix, launching new models such as T1-Vision and Hunchun Voice. TurboS has performed excellently in international evaluations, making new progress in multimodal technology, while also promoting the development of open-source technology.
【AiBase Summary:】
🚀 Hunchun has introduced T1-Vision and Hunchun Voice, enhancing visual and voice processing capabilities.
🏆 TurboS model ranks among the top eight globally, with scientific reasoning ability increased by over 10%.
🌐 Tencent Hunchun has open-sourced multimodal technology, supporting industry innovation and partner development.
3. Meitu reaches strategic cooperation with Alibaba, receiving $250 million investment
Meitu Company has reached a strategic cooperation with Alibaba, which will invest $250 million in Meitu through convertible bonds. Both parties will conduct in-depth cooperation in e-commerce, AI technology, cloud computing, and other fields.
【AiBase Summary:】
Alibaba invests $250 million in Meitu via convertible bonds, helping its AI technology and e-commerce business development.
Meitu and Alibaba will deepen their cooperation in the development of e-commerce tools and procurement of cloud services to enhance merchant operation efficiency.
The cooperation requires government approval conditions and aims to promote technological innovation and business growth.
4. DouBao Voice Podcast Model Released - Will be available on DouBao APP, PC end, and buttons, etc.
Vesuvius Engine's DouBao Voice Podcast Model is built through a streaming model construction, enabling rapid conversion from text to podcast, providing an efficient and interactive creation experience.
【AiBase Summary:】
🌟 Conversations are natural and smooth, simulating real-person podcast oral habits to enhance listening realism.
⚡ End-to-end convenient chain, from creation to finished product in seconds, greatly improving efficiency.
🔍 Supports deep search and long-text-to-podcast, always following hot topics and enriching content.
5. Baidu Wenxin X1 Turbo shines, winning the highest rating from the China Academy of Information and Communications Technology!
Baidu Wenxin large model X1 Turbo has received the highest "4+ level" rating in the evaluation by the China Academy of Information and Communications Technology, demonstrating outstanding technical strength.
【AiBase Summary:】
🚀 Wenxin X1 Turbo stands out in 24 capability assessments, scoring full marks in 16 capabilities, particularly excelling in logical reasoning and code reasoning.
🌟 Based on the upgrade of Wenxin large model 4.5, X1 Turbo performs better with lower costs and recognized reliability and safety.
🌐 Baidu continues to innovate in AI technology research and development, and the success of X1 Turbo marks an important breakthrough in domestic AI technology.
6. Google releases Flow AI editing tool: Veo3 and Imagen4 driven, lens control and scene expansion
At the 2025 Google I/O Developer Conference, Google launched a new AI editing tool called Flow designed for filmmakers. It integrates three major AI models, Veo3, Imagen4, and Gemini, simplifying the video creation process through natural language prompts and modular workflows, providing professional and beginner creators with cinematic visual storytelling capabilities.
【AiBase Summary:】
🎥 Flow integrates Veo3, Imagen4, and Gemini, generating high-quality AI video clips through simple text prompts, supporting character dialogues and ambient sound effects.
🎥 Scenebuilder and Camera Controls functions give creators seamless editing and lens control capabilities.
🎥 Built-in Asset Management and Flow TV platforms optimize material management and provide learning resources, helping develop the creative community.
For more details: https://labs.google/flow/about
7. From Text to Movie! Google Veo3震撼发布: 4K resolution, audio-video integration
Google's released Veo3 video generation model not only supports audio-video integration but also features ultra-high resolution and diverse style support, providing powerful tools for creators while emphasizing safety and responsibility.
【AiBase Summary:】
🎥 Veo3 can generate high-definition videos and matching audio simultaneously, simplifying the creation process, with excellent physical simulation and lip synchronization effects.
🌍 Supports 4K resolution, capturing fine details, and adapting to various styles, suitable for professional and educational use.
🔒 Access is limited to specific users, integrating security features like digital watermarks and filters, with future expanded application scenarios.
8. Google Imagen 4 Image Model震撼发布: 2K resolution and precise text rendering reshaping AI creation
Google has released the latest generation image generation model Imagen4, featuring ultra-high resolution, precise text rendering, and multiple style support, providing powerful tools for creators.
【AiBase Summary:】
✨ Ultra-high resolution and detail expression: supports 2K resolution, significantly improves detail capture, meeting professional design needs.
🌟 Improved text rendering and style diversity: generates clear text, supports various art styles, expanding creator expression space.
🚀 Fast generation mode and ecosystem integration: speed increases by 10 times, integrated into multiple platforms, ensuring content traceability and reducing misuse risks.
9. Google Gemini 2.5 Pro Deep Think release: Parallel inference reshapes complex problem-solving for AI
In 2025 Google I/O Developer Conference, Google DeepMind launched the Gemini2.5Pro Deep Think mode, using parallel inference technology to significantly improve mathematical, coding, and multimodal reasoning capabilities, allowing developers to flexibly adjust inference quality and cost.
【AiBase Summary:】
💎 Deep Think mode significantly improves performance in complex tasks through parallel inference technology, such as mathematics, coding, and multimodal reasoning.
⚙️ Introduces thinking budget functionality, allowing developers to balance response quality, latency, and cost, adapting to different scenario needs.
🔒 Currently only available to trusted testers, and will gradually expand to broader application areas in the coming weeks.
10. Google's Virtual Try-On Room goes live! Upload photos and become a fashion model instantly, AI one-click price comparison + automatic order placement
Google launched a new AI virtual try-on tool at I/O 2025. Users can quickly generate realistic outfit effect pictures by uploading photos, while supporting multi-condition searches, price comparisons, and automatic order placements, greatly enhancing personalization and convenience in online shopping.
【AiBase Summary:】
✨ Use AI virtual try-on tools, upload photos, and within seconds, generate realistic outfit effect pictures, experiencing high personalization.
🔍 Integrates multi-condition searches, real-time price comparisons, and automatic order placements, simplifying the shopping process and enhancing shopping efficiency.
🌍 Based on Gemini2.5 and Shopping Graph technology, accurately captures clothing details, supporting various body types and brands.
11. Google’s AI Code Assistant Jules enters global public testing: Powered by Gemini 2.5 Pro, challenging OpenAI’s Codex
Google's AI code assistant Jules has entered the global public testing phase, based on the Gemini 2.5 Pro model. Through asynchronous task processing and GitHub integration, it enhances developer efficiency. Transparent workflows and free quotas attract users, with great potential for future expansion.
【AiBase Summary:】
🌟 Asynchronous task processing and seamless integration with GitHub automate bug fixes, writing tests, etc.
📝 Transparent workflow, generating detailed plans and code change differences, ensuring code quality and team standards.
🚀 Supports Python and JavaScript, with future expansion to more languages, challenging OpenAI Codex and GitHub Copilot.
For more details: https://jules.google/
12. Apple plans to open up AI models to third-party developers in iOS 19
Apple plans to open up its AI models to third-party developers in iOS 19, promoting application innovation and enhancing hardware appeal. Initially, it will open small AI models that can run locally on devices, allowing developers to create new features using the new toolkit and Apple Intelligence.
【AiBase Summary:】
🍎 Apple plans to open AI models in iOS 19, driving application innovation and enhancing hardware appeal.
💻 Developers can create new features using Apple Intelligence, initially only opening local small AI models.
🌟 Apple's goal is to simplify the development process, transforming the operating system into a core platform for AI software.
13. JD Cloud announces five AI marketing products free of charge, helping merchants improve sales efficiency
JD has launched five AI marketing products, including digital person live streaming, merchant customer service intelligent body, product image generation service, video mixing platform, etc., provided free of charge to merchants to enhance sales efficiency and operational capabilities.
【AiBase Summary:】
🌟 JD offers five AI marketing products free of charge, helping merchants improve sales efficiency.
🤖 Digital person live streaming can sell 24/7, reducing live streaming costs and increasing conversion rates.
📸 Jingdian platform and video mixing tools help merchants optimize graphic and video marketing.
14. ByteDance open-sources multimodal model BAGEL: New breakthroughs in text-to-image generation and editing
ByteDance has released an open-source multimodal foundation model named BAGEL, featuring 7 billion active parameters and a total parameter count of 14 billion. The model performs exceptionally well in multimodal understanding, text-to-image generation, and image editing, employing a hybrid transformer expert (MoT) architecture and achieving continuous performance improvement through multimodal pre-training.
【AiBase Summary:】
🌟 BAGEL is an open-source multimodal foundation model with 7 billion active parameters, surpassing multiple standard benchmark tests.
🖼️ The model performs excellently in image generation and editing tasks, capable of free-form image editing and world navigation.
📈 Through multimodal pre-training, BAGEL demonstrates continuous performance improvement, adapting to complex multimodal reasoning tasks.
For more details: https://huggingface.co/ByteDance-Seed/BAGEL-7B-MoT
15. NVIDIA introduces new Cosmos-Reason1 model, enhancing AI understanding of the physical world
NVIDIA’s released Cosmos-Reason1 series models improve AI capabilities in physical common sense and embodied reasoning through physical AI supervised fine-tuning and reinforcement learning, bringing new possibilities to robotics and autonomous driving fields.