Welcome to the "AI Daily" section! Here 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, helping you understand technology trends and learn about innovative AI product applications.
Fresh AI products click to learn more:https://app.aibase.com/zh
1. Alibaba Quark launched the AI creation platform "Zhaodian", Tongyi Wan2.5 generates video for 7 days free
Alibaba Quark launched the AI creation platform "Zhaodian", integrating Tongyi Wan2.5 and Midjourney V7, supporting synchronized audio-visual video generation and advanced image creation features, while providing convenient AI creation tools for ordinary users.
【AiBase Summary:】
🎥 Integrated with Tongyi Wan2.5, supports synchronized audio-visual video generation.
🎨 Introduces Midjourney V7, provides multiple image references and style codes as advanced functions.
📱 Provides a lightweight creation experience, meeting the daily photo editing and creative needs of ordinary users.
2. Wan2.5-Preview released, achieving multimodal input and movie-level video synchronization generation
The release of Wan2.5-Preview marks a new stage in AI visual generation technology, its powerful multimodal capabilities and precise control functions will provide developers and creators with unprecedented tools.
【AiBase Summary:】
🎥 Video generation supports high-fidelity, consistent audio-visual synchronization, can generate multiple voices, sound effects and BGM.
🎨 Image generation capability improved, can generate realistic images, diverse artistic styles and professional charts.
🔄 Supports dialogue-based, instruction-based image editing, achieving pixel-level precision operations.
3. Keling AI launched the latest video generation model Keling 2.5 Turbo and announced a price reduction of nearly 30% compared to the previous version
The article introduces Keling AI's latest video generation model Keling 2.5 Turbo and mentions that its price has dropped by nearly 30% compared to the previous version. At the same time, the article also mentions the pricing strategies of other manufacturers such as Hai Luo AI and Baidu in the field of large-scale video generation models, as well as the views of industry experts on video generation models.
【AiBase Summary:】
🎥 Keling AI launched the latest video generation model Keling 2.5 Turbo and announced a price drop of nearly 30%.
💰 Keling AI's pricing system includes multiple levels, with the basic membership only costing 66 yuan, while the highest level membership costs 1314 yuan per month.
📈 Keling AI is the only large model manufacturer that publicly disclosed revenue, with second-quarter revenue exceeding 250 million yuan.
4. Alibaba Tongyi launched Qwen3-ASR-Toolkit, achieving a breakthrough in audio-video transcription
The Alibaba Tongyi Qwen team launched Qwen3-ASR-Toolkit, an open-source Python command-line tool that can achieve hour-level audio-video transcription. The tool is based on the latest Qwen3-ASR-Flash model, supports various audio-video formats, and has intelligent silent segmentation and multi-threaded parallel upload functions, significantly improving transcription efficiency and accuracy.
【AiBase Summary:】
🎧 Breaks the three-minute limit, supports hour-level audio-video transcription.
🧠 Based on the Qwen3-ASR-Flash model, ensuring high accuracy in speech recognition.
🔄 Supports various audio-video formats, enhancing transcription flexibility and efficiency.
Details link: https://github.com/QwenLM/Qwen3-ASR-Toolkit
5. Google Photos AI Editing Revolution: Android Users Can Edit Photos with Just Words, Pixel 10 Exclusive Features Fully Opened
The article introduces Google's new AI editing feature, which allows Android users to easily edit photos through natural language interaction, solving the complexity of traditional photo editing. This feature is now available to all Android users from Pixel 10 devices, enhancing mobile photography experiences and supporting advanced operations such as light adjustment and background removal.
【AiBase Summary:】
🤖 Google launched an AI editing feature, allowing users to easily edit photos through natural language instructions.
📷 The feature covers a wide range, including advanced operations such as light adjustment, background removal, and old photo restoration.
🌐 Google promotes AI transparency, supports C2PA content credentials to maintain information authenticity.
6. Google Launches Mixboard: An AI Tool That Makes Creative Design Easier
Google launched an experimental tool called Mixboard, aimed at helping users quickly create mood boards, explore and realize design inspiration. The tool supports templates and text prompts, and allows users to generate visual effects by uploading images or using natural language descriptions, suitable for home decoration, event planning, and other scenarios. In addition, it supports one-click regeneration and image editing features, improving design efficiency.
【AiBase Summary:】
🎨 Mixboard is a new AI tool launched by Google, helping users quickly create personalized mood boards, supporting templates and text prompts.
🖌️ Users can generate visual content by uploading images or using natural language, suitable for home decoration and event planning.
🔄 The tool supports one-click regeneration and image editing, enhancing the flexibility and convenience of the creative process.
Details link: https://labs.google.com/mixboard/welcome
7. Qwen Launches New AI Model Qwen3-Max: Not Only Can It Write Code, But Also Think Autonomously
The Qwen team launched a new AI model Qwen3-Max, which performs well in multiple benchmark tests, especially in code generation and agent capabilities. Qwen3-Max has a large parameter scale and uses an advanced MoE structure design, improving training efficiency and stability. In addition, Qwen3-Max-Instruct and Qwen3-Max-Thinking also show excellent performance, providing developers with powerful tools.
【AiBase Summary:】
🧠 Qwen3-Max model has over 1 trillion parameters, pre-training used 36 trillion tokens.
💻 Qwen3-Max-Instruct achieved excellent results in programming benchmarks, demonstrating its ability in practical programming.
🧮 Qwen3-Max-Thinking scored full marks in mathematical reasoning tests, proving its strong reasoning ability.
8. Figma MCP Server Upgraded! One-Click Design to Code, Designers and Front-End Developers Bid Farewell to the "Translation" Era
Figma's MCP server has completely changed the way design and development collaborate, allowing AI models to directly read semantic layer information of design files through standardized protocols without relying on clients. This not only improves development efficiency but also realizes deep integration between design systems and front-end development.
【AiBase Summary:】
✅ Remote access function allows AI models to directly read semantic layer information of design files without screenshots or manual descriptions.
🔄 Seamlessly integrated with Figma Make, supports extracting resources from Make files and converting them into code-level assets.
⚡ Code Connect UI component mapping optimization, improves the accuracy of AI-generated code, shortening development time by 60%-80%.
Details link: https://127.0.1:3845/sse
