Welcome to the "AI Daily" section! This is your guide to exploring the world of artificial intelligence every day. Each day, we present you with the latest content in the AI field, focusing on developers, helping you understand technology trends and innovative AI product applications.
Fresh AI products Click to learn more: https://app.aibase.com/zh
1. Kunlun Wanshi's AI video product SkyReels will launch a new version in early November
Kunlun Wanshi's SkyReels AI video product is about to release a new version, expected to be officially launched in early November. The product relies on the company's long-term technical accumulation in the AI field, aiming to promote the popularization of AI video creation, achieving efficiency and accessibility in professional video creation.
【AiBase Summary:】
🚀 SkyReels is an important part of Kunlun Wanshi's "All in AGI and AIGC" strategy, aiming to drive the AI video productivity revolution.
🎥 SkyReels-V1 model achieves precise control of 33 micro-expressions and over 400 action postures, reaching SOTA levels in the open-source field.
🔊 SkyReels-A3 audio-driven model supports full-modal audio-driven digital human creation for any length, enhancing creative flexibility.
More details: https://www.skyreels.ai/home
2. LiblibAI completes $130 million funding, becoming the largest single investment in the domestic AI application field
LiblibAI successfully completed its $130 million Series B funding, becoming the largest single investment in the domestic AI application field. The company, leveraging its leading position in multimodal models and creation communities, has rapidly grown into the largest AI application platform in China and plans to accelerate its global expansion.
【AiBase Summary:】
💰 LiblibAI completes $130 million Series B funding, becoming the largest single investment in the domestic AI application field.
🖼️ The platform integrates various multimodal capabilities such as images, videos, 3D, and LoRA training, covering the entire AI workflow.
🌐 The company plans to accelerate global expansion, building a global creator co-creation multimodal content ecosystem.
3. Alibaba's 'C Plan' debut: Qwen-powered Quark Chat Assistant launches, seizing the consumer AI entry point
Alibaba launched the first product of its 'C Plan'—Quark Chat Assistant, using the Qwen model to create an AI entry point for young users, integrating multiple functions and achieving the integration of search and conversation experience, marking a significant move in Alibaba's consumer application ecosystem.
【AiBase Summary:】
🧠 Quark Chat Assistant uses the latest closed-source Qwen model, integrating functions such as Q&A, search, and photo-based question solving.
🔍 Quark combines the advantages of photo-based question solving and information retrieval to achieve the integration of search and conversation experiences.
🚀 Alibaba Group has elevated the priority of the 'C Plan' to the highest level, aiming to seize the consumer AI entry point.
4. Doubao Video Generation Model Seedance 1.0pro Launches Start and End Frame Capabilities
The start and end frame capabilities of the Doubao video generation model Seedance 1.0pro have been launched, marking a significant progress in controllability and consistency in AI video creation. The model has technical advantages such as maintaining the consistency of main subjects in complex scenes, ensuring physical rationality in large movements, and intelligent reasoning of video rhythm, which can enhance the effect of character following in generated videos, achieve accurate narrative guidance, and generate more immersive and expressive video works.
【AiBase Summary:】
🧠 Seedance 1.0pro enhances structured understanding of "narrative subjects," locking in core character features.
🏃♂️ In large movement scenes, the model can accurately capture human motion trajectories, keeping actions coherent and reasonable.
🌊 Deep semantic understanding allows the overall video rhythm to be natural and smooth, complying with physical logic.
5. Vidu Q2 Reference Student MaaS Platform Fully Opens API Access
The full opening of the Vidu Q2 Reference Student video large model API marks a key leap from 'usable' to 'precision engineering' in AI video generation technology. It shows unique value in areas such as advertising and product display, enabling precise restoration of product details and infusion of emotional expression, thus enhancing brand favorability and user conversion.
【AiBase Summary:】
🎥 Vidu Q2 sets a new benchmark in the realism of video generation, injecting real emotions into digital characters through subtle facial expressions.
🔄 New video extension function supports up to 5-minute video generation and adds sound effect selection, providing more possibilities for enterprise customers.
🎨 Offers holiday special effects templates, such as Halloween template collections, allowing users to easily create creative dynamic videos.
More details: https://platform.vidu.cn
6. Sora Roadmap Major Update: Launches "Guest Role" Feature, Android Version Coming Soon
Sora's roadmap update demonstrates significant progress in the AI video generation field, including the guest role feature, the release of the Android version, and optimization of the social experience, further promoting the openness and diversity of AI video creation.
【AiBase Summary:】
🎭 The guest role feature allows pets and plush toys to participate in video creation.
📱 The Android version is coming soon, broadening the user base.
💬 Enhanced social experience, supporting private channels and community interaction.
7. Hailuo 2.3 to Launch: Next-Generation AI Video Model Exceeding Veo, Realism Upgraded
Hailuo 2.3 is the latest version of MiniMax's flagship video generation model, achieving significant breakthroughs in realism, accuracy, and style diversity, and is considered a major leap in the AI video generation field.
【AiBase Summary:】
🎥 Hailuo 2.3 shows higher fidelity in motion capture and facial expressions.
🎨 The new version enhances ultra-realistic character animation and micro-expression capture capabilities.
🎬 Improves motion stability and physical consistency, eliminating flickering and inconsistent actions.
8. Google Launches "Google Skills" Platform, Opening Internal AI Knowledge to the Public Free of Charge
Google launched a new learning platform called 'Google Skills', opening the essence of DeepMind, Google Cloud, Gemini AI model development teams, and Google Education to the global public. The platform offers a zero-barrier learning experience, including real AI research courses, over 700 hands-on labs, AI coding assistants, and career pathways, aiming to bridge the skills gap in the AI field.
【AiBase Summary:】
🌍 The Google Skills platform systematically opens internal Google AI knowledge for the first time, helping global learners master cutting-edge technologies.
💻 Provides hands-on labs in real cloud environments, allowing students to write code and receive real-time feedback.
🤝 After completing the course, you can obtain official certification and establish direct employment channels with over 150 companies, enhancing career competitiveness.
More details: https://www.skills.google/
9. ByteDance Seed Team Announces Launch of 3D Generation Large Model Seed3D1.0
ByteDance's Seed team launched Seed3D1.0, an innovative 3D generation large model that can generate high-quality simulation-level 3D models from a single image, including detailed geometry, realistic textures, and PBR materials. The model shows significant advantages in multiple evaluations, especially in geometric generation and texture map generation. Additionally, Seed3D1.0 supports building complete 3D scenes and can be seamlessly imported into simulation engines, providing strong support for the development of embodied intelligence through powerful world simulators.
【AiBase Summary:】
🧠 Seed3D1.0 uses a Diffusion Transformer architecture to quickly generate simulation-level 3D models from a single image.
🧱 Seed3D1.0 excels in geometric generation, accurately constructing structural details and ensuring physical integrity.
🕹️ Seed3D1.0 supports generating complete 3D scenes and can be seamlessly imported into simulation engines like Isaac Sim, providing diverse operational scenarios for robot training.
More details: https://seed.bytedance.com/seed3d
