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.
Fresh AI products Click to learn more:https://app.aibase.com/zh
1、Tongyi Lab of Alibaba releases Wan2.7-Video video generation model
The Wan2.7-Video video generation model launched by Tongyi Lab of Alibaba solves the issues of unprofessional content and difficult modifications in AI video creation through advanced technology, providing creators with higher freedom and flexibility.
【AiBase Summary:】
🎥 Supports full-modal input including text, images, videos, and audio, achieving precise control.
🖌️ Users can precisely adjust video elements such as removing people or replacing objects via instructions.
🔄 Provides creative replication and plot continuation features to enhance narrative fluency.
2、Red Fruit Short Drama removes AI drama "Peach Hairpin": Producer involved in "face theft" infringement suspended
Red Fruit Short Drama removed the AI short drama "Peach Hairpin" due to unauthorized use of a blogger's facial image, and the producer was suspended from uploading for 15 days. The incident sparked discussions on the legal boundaries of AI image rights infringement, highlighting the legal risks and industry governance challenges of generative AI in film and television production.
【AiBase Summary:】
🎭 The AI short drama "Peach Hairpin" was removed due to unauthorized use of a blogger's facial image.
🚫 The producer was suspended from uploading for 15 days due to inability to prove the compliance of the materials.
⚖️ The incident triggered discussions on the legal boundaries of AI image rights infringement.
3、Tencent Cloud launches “Lobster” Memory Service Agent Memory, enhancing AI assistant intelligence level
Tencent Cloud launched the “Lobster” memory service - TencentDB Agent Memory, adding powerful long-term memory capabilities to the OpenClaw system. This innovative memory engine uses a four-layer progressive memory system, allowing the AI assistant to better understand and remember user needs. The introduction of Agent Memory service has increased the answer accuracy of OpenClaw to 76.10%, nearly 59% higher than the previous memory system. Users can experience that Lobster not only handles simple daily conversations but also maintains information coherence in complex projects, avoiding the "amnesia" phenomenon caused by lack of long-term memory.
【AiBase Summary:】
🧠 The Agent Memory service improved the answer accuracy of OpenClaw to 76.10%.
💡 A four-layer memory architecture ensures the AI assistant can accurately remember user needs and preferences.
🔧 Users can easily enable the Agent Memory plugin with a single operation; the enterprise version will be launched soon.
4、Qwen 3.6 officially released: 1 million long context, competing with Claude Code
The release of Qwen 3.6-Plus marks a major breakthrough in the programming field of domestic large models, with its performance and functions reaching international advanced levels, providing global developers with stronger tools.
【AiBase Summary:】
🧠 Domestic model performance has taken a big leap, directly competing with the programming benchmarks of the Claude series.
💻 Supports long-range planning tasks, achieving high-quality code repair and automation.
🌐 Strong ecological compatibility, supports million-level context and mainstream agent frameworks.
5、Broadcasting industry organization speaks out: Strictly prohibit AI actors' face swapping and voiceprint cloning
The Chinese Radio and Television Association of Artists issued a statement clearly prohibiting AI face swapping and voiceprint cloning, and requiring platforms to strengthen review and verification mechanisms to protect the legitimate rights and interests of actors.
【AiBase Summary:】
🚫 Prohibits unauthorized use of actor's image and audio data.
🔍 Platforms need to establish verification mechanisms to check and remove infringing works.
⚖️ The committee will initiate routine monitoring and pursue legal responsibility for infringing parties.
6、ElevenLabs launches iOS app ElevenMusic, officially entering the AI music creation market
The ElevenMusic application launched by ElevenLabs marks its official entry into the AI music creation and social discovery field, directly competing with platforms like Suno and Udio. The app allows users to generate customized music through natural language prompts, and supports adjusting track length, lyric switches, and artistic styles. In addition, it integrates streaming social attributes, offering real-time radio, preset albums, and curated playlists based on emotional dimensions.
【AiBase Summary:】
🚀 ElevenMusic allows users to generate customized music through natural language prompts.
📱 The app deeply integrates streaming social attributes, offering real-time radio and curated playlists.
💰 It adopts a "free + subscription" model, with professional versions offering higher creation quotas and storage space.
7、Google officially releases Gemma4 open-source large model: includes four specifications, 31B version ranks third globally in open-source lists
Google officially released the open-source large model Gemma4, which includes four specifications, with the 31B version ranking third globally in open-source lists, demonstrating its deep roots and technical strength in the open-source ecosystem.
【AiBase Summary:】
🧠 The Gemma4 series includes efficient versions E2B (2.3B) and E4B (4.5B), as well as high-performance versions 26B MoE and 31B dense models.
🖼️ Supports multimodal input, including images, videos, and voice, enabling real-time voice understanding at the edge.
⚙️ Reduces the threshold for local deployment, supporting consumer-grade GPUs and mobile devices, improving privacy and localized application capabilities.
8、Xiaomi MiMo large model launches Token subscription plan: four tiers cover all modalities, starting at 39 yuan per month
Xiaomi officially launched the first Token Plan subscription package for its self-developed MiMo large model, marking its entry into the stage of large-scale delivery in AI commercialization. The package covers multiple models and modalities, offering transparent billing methods, and driving the transition of domestic large models from "technical storytelling" to "operational drive" through strong technological power and business model innovation.
【AiBase Summary:】
🧠 MiMo large model introduces four-tier Token Plan subscription packages, priced from 39 yuan to 659 yuan per month, achieving transparent billing for multi-model and multi-modal calls.
🚀 MiMo-V2-Pro, with trillions of parameters and millions of context windows, ranks fifth in the global Model Rank in the Text Arena double-blind test.
🌐 Xiaomi transforms unpredictable R&D costs into budgetable and scalable product services through the "monthly subscription" model, building a complete closed loop from technology research and development to commercial operations.
Details: https://mimo.mi.com/
