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. Surpassing Seedance 2.0! Happy Horse, a mysterious dark horse in the AI video field, has emerged suddenly
The article reports on the rise of the new model Happy Horse in the field of AI video generation. Its performance surpasses existing leading models like Seedance 2.0 and it has multi-modal video creation capabilities. The emergence of this model has sparked heated discussions in the global AI community and is considered an important sign of industry technological iteration.
【AiBase Summary:】
🎥 The Happy Horse model performs well in text-to-video generation, with superior image consistency, detail restoration, and natural motion.
🔊 The model supports audio generation, achieving true multi-modal video creation capabilities.
🔍 The origin of this model remains mysterious, possibly from an Asian AI laboratory, and is seen by the industry as a potential rival to Seedance.
2. 15 seconds, 1080P Audio-Visual Synchronization! AiShi Technology PixVerse C1 Launches: A Professional Large Model for the Film Industry Officially "Explosive"
AiShi Technology launched a professional large model PixVerse C1 for the film industry. This model not only supports 1080P high-definition video generation of 15 seconds but also has core capabilities such as automatic scene segmentation, control of the first and last frames, and native audio-visual synchronization, providing a new solution for film production.
【AiBase Summary:】
🎥 Leap from "Single Shot" to "Automatic Scene Segmentation"
🎬 High Definition Video Output: Supports up to 15 seconds of 1080P high-definition video generation.
🎙️ Native Audio-Visual Synchronization: Video output comes with built-in sound effects, achieving true audio-visual integration.
3. Tencent officially released “Lobster” QBotClaw: The first AI browser in China that supports free configuration of major large model APIs
Tencent officially released the first AI browser in China that supports free configuration of major large model APIs, “Lobster” QBotClaw. This marks the evolution of browsers from simple web access tools into AI assistants with full-scene perception and complex task execution capabilities. The core highlight of this product is its high openness, allowing users to freely configure API keys for major domestic large models, breaking the previous limitation of AI browsers being tightly bound to a single model. The initial Mac version has been deeply integrated with QQ Browser Skill, and the Windows version is expected to be released soon, aiming to reduce the barriers to using large models and achieve the industry's goal of "everyone can use Lobster." Technically, QBotClaw uses the X5use high-precision recognition technology to give the browser keen visual perception, combined with context awareness, so users can simply instruct the browser to autonomously perform complex cross-page tasks. In addition to fully compatible with the OpenClaw skill ecosystem, QBotClaw also breaks down the barriers between mobile and desktop devices, allowing users to remotely control the computer via WeChat scan binding with Clawbot. Currently, the global large model field is at a critical stage of transitioning from "dialogue interaction" to "intelligent agent execution." Tencent's release of QBotClaw not only restructures the core traffic entrance of browsers through AI, but also aims to build a comprehensive AI application ecosystem through its open API configuration mechanism.
【AiBase Summary:】
🧠 Supports users to freely configure API keys for major domestic large models, breaking the limitations of AI browsers being tied to a single model.
💻 The initial Mac version has been deeply integrated with QQ Browser Skills, and the Windows version will be launched soon, lowering the threshold for using large models.
🔄 Breaks down the barriers between mobile and desktop, allowing remote control of the computer via WeChat scanning and binding Clawbot.
4. 360 Creates "Shaxiu" APP: A Digital Community for Observing AI Agent Conflict
360 Group's "Shaxiu" APP, centered around AI agents, offers a new social experience, allowing users to transition from "users" to "observers," and attempts to build a complete AI-native application cluster.
【AiBase Summary:】
🧠 Based on the OpenClaw ecosystem, the AI "Shrimp" has independent thinking and self-awareness.
👀 Users participate in the interaction between AI agents as observers, experiencing a social battlefield.
🌐 360 Group is building an AI agent ecosystem through "Shaxiu," creating a complete AI-native application cluster.
5. DeepSeek V4 Gray Scale Test Exposed: New Visual Version and "Expert Mode" Appear
DeepSeek V4 is currently undergoing a gray scale test, featuring a new functional architecture including a fast version, expert version, and visual version, while emphasizing compatibility with domestic chips, demonstrating strong technical capabilities and market competitiveness.
【AiBase Summary:】
🧠 The new generation of models has made significant upgrades in underlying architecture, interaction logic, and multimodal capabilities.
💻 DeepSeek V4 prioritizes compatibility with domestic AI chips, refusing to rely on foreign chip vendors.
🚀 It is expected to launch an AI programming special edition, challenging products from Anthropic and OpenAI.
6. Mythos, a Powerful AI Model from Anthropic, Is Limited to Security Partners for Trial Use
Anthropic has launched its new AI model, Mythos, which has powerful programming and reasoning capabilities, capable of discovering thousands of zero-day vulnerabilities, especially excelling in uncovering historical vulnerabilities. However, due to potential security risks, the model is limited to specific security partners for trial use, and the company is discussing its application in national key systems with U.S. federal officials.
【AiBase Summary:】
🌟 Mythos is the latest AI model introduced by Anthropic, available only to a few security partners for trial use.
🔍 The model discovered thousands of zero-day vulnerabilities during testing, showcasing its potential in uncovering historical vulnerabilities.
⚠️ The company is discussing its application in national key systems with U.S. federal officials, while facing legal disputes.
7. Digital Family Members Board! Doubao Large Model First Launched on BYD Zhijing E7: Smart Cabin Enters the "Human-like" Era
The article introduces how SAIC General Motors applied the Doubao large model to the BYD Zhijing E7, marking a new stage in smart cabins moving from "command interaction" to "semantic understanding."
【AiBase Summary:】
🧠 Emotion Recognition: The system supports over 20 emotional expressions and can identify users' joy, anger, sorrow, and happiness through tone and speech rate.
🔄 Context Memory: It has continuous dialogue capabilities, automatically filtering out locations that meet specific needs and planning routes.
🔒 Safety Isolation: Built a strict three-layer safety isolation wall to ensure the security of AI interactions.
8. Zhipei Releases GLM-5.1: Leading in SWE-bench Score, Model Price Increased by 10%
Zhipei released the new large model GLM-5.1, which performed excellently in the SWE-bench Pro benchmark test, and aligned its pricing with international top manufacturers, marking the industry competition focus shifting towards performance premium.
【AiBase Summary:】
🧠 GLM-5.1 set a new world record in the SWE-bench Pro benchmark test, surpassing Claude4.6Opus and ranking first among domestic and open-source models.
💰 Zhipei's entire GLM series increased prices by 10%, and the cache hit token price in core scenarios like coding has basically been anchored to Claude3.5Sonnet from Anthropic.
🚀 GLM-5.1 has strong autonomous engineering capabilities, able to work continuously for up to 8 hours in a single task, independently completing planning, execution, and testing.
