Search AI Products and News

In the article "From AI to IA, Whoever Masters Agents Masters the World", we explored how the core driving force behind the explosive development of the AI era is the paradigm shift in industrial applications brought about by multi-agent collaboration. From intelligent question answering to task execution, Agents are bringing AI to real-world applications in vertical fields, creating new business models. Multi-agent collaboration (InterAgent) should follow a specific standard framework to achieve maximum scalability and interoperability. Based on our theoretical exploration and practical experience, we attempt to outline this framework here.

Baidu announced today the full launch of Miaoda, its domestically pioneered conversational application development platform. This platform allows users to describe their needs using natural language, enabling the system to automatically generate application code with complete functionality, marking a significant breakthrough in the field of no-code development tools.

The internal testing of the AI agent Manus has sparked significant excitement, with users praising its groundbreaking capabilities and demanding access. Its innovative features are revolutionizing the traditional virtual assistant experience.

Generating charts that accurately reflect complex data remains a subtle challenge in today's data visualization landscape. Charts require not only precise layout, color, and text placement, but also the translation of these visual details into code to reproduce the intended design. However, traditional methods often rely on directly prompting vision-language models (VLMs), such as GPT-4V, which frequently struggle to translate complex visual elements into grammatically correct Python code. Even minor errors can result in charts failing to meet their design objectives.

At the 2025 Global Developer Pioneer Conference, SenseTime announced the launch of its open-source low-code platform, LazyLLM, aimed at lowering the barrier to entry for AI application development. Developers can now build complex multi-agent large model applications with approximately 10 lines of code. LazyLLM not only helps users rapidly develop AI applications but also simplifies the development process through intuitive operations, making it as easy as building with blocks. The name "LazyLLM" reflects SenseTime's goal of enabling effortless AI application development.

With the advancement of artificial intelligence technology, multi-agent systems are increasingly capable of handling complex tasks across various fields. These systems are composed of multiple specialized agents that collaborate to leverage their individual strengths and achieve common goals. Such collaboration excels in areas like complex reasoning, programming, drug discovery, and safety assurance, as the structured interactions between agents not only enhance problem-solving efficiency but also allow for mutual correction, thereby improving each agent's output. Research indicates that this collaborative approach often outperforms in tasks that require rigorous reasoning or fact verification.

H2O.ai recently announced the launch of its new multi-agent platform h2oGPTe, which combines generative and predictive AI models to provide businesses with more consistent responses. Sri Ambati, founder and CEO of H2O.ai, stated that a major demand from enterprises for AI agents is to maintain response consistency. The h2oGPTe platform utilizes H2O.ai's proprietary models Mississippi and Danube, while also allowing access to other models.

On the occasion of Vidu's launch anniversary, Shengshu Technology is proud to announce the release of the new version Vidu 1.5, which achieves world-leading breakthroughs, especially in understanding diverse inputs and overcoming the 'consistency' challenge. The launch of Vidu 1.5 marks a new 'contextual' era for visual models, accelerating the arrival of Artificial General Intelligence (AGI). From its global debut, Vidu has demonstrated the ability to generate character consistency.

At the 2024 Baidu World Conference, Baidu introduced an innovative no-code tool called 'MiaoDa', personally launched by Baidu founder Robin Li. This tool integrates three core functionalities: no-code programming, multi-agent collaboration, and multi-tool invocation, and was hailed by Robin Li as 'the most complex multi-agent collaboration tool in human history.'

Recently, OpenAI announced that it is recruiting machine learning engineers for its newly established multi-agent research team. The news was shared by OpenAI's research scientist Noam Brown via social media. He stated, 'We believe that multi-agent systems are an important pathway to enhance AI reasoning capabilities, and previous experience in multi-agent systems is not a prerequisite.' Brown also encouraged professionals interested in researching this field to apply and join the team.

Text-to-SQL technology allows ordinary users to query databases easily through natural language, but accurately converting complex database structures into SQL commands poses challenges. A research team from South China University of Technology and Tsinghua University has proposed a new method—MAG-SQL (Multi-Agent Generative Model), aimed at enhancing the accuracy of this process. MAG-SQL achieves this through the collaboration of multiple agents, including a 'Soft Pattern Linker', a 'Goal-Condition Decomposer', a 'Sub-SQL Generator', and a 'Sub-SQL Refiner'.

Recently, Tencent AI Lab has introduced a new multi-agent framework called "TRANSAGENTS", specifically designed for translating ultra-long literary content. This framework is set to revolutionize the field of literary translation, offering readers enhanced cross-cultural reading experiences.This virtual translation publishing company simulates the real translation process, including roles such as senior editors and junior editors, to achieve more precise and efficient literary translation. It is

Researchers have open-sourced a project called AgentTuning on GitHub, which provides a new method for fine-tuning language models. AgentTuning trains and adjusts language models through interaction trajectories from multiple agent tasks to adapt to different tasks and scenarios. This method can enhance the performance and generalization ability of language models while reducing the workload of manual tuning. AgentTuning has been validated in various natural language processing tasks, including dialogue generation, question answering systems, and summarization.

The Waymo research team has launched MotionLM for predicting the behavior of road traffic participants. Unlike traditional methods, MotionLM utilizes language modeling and does not require complex optimization programs. This method demonstrates exceptional performance in predicting the behaviors of road participants and is applicable to autonomous vehicles. MotionLM can simulate temporal causal relationships, enhancing prediction accuracy. This innovative approach is expected to promote advancements in the autonomous driving field.

["MIT research found that multi-agent debates can significantly improve the accuracy and reasoning abilities of AI models.", "The debate method reduces hallucinations and helps improve the model's responsiveness.", "Multiple AI systems debating answers to problems are more accurate than a single system.", "This method can be used for collaborative work among different language models to improve output.", "These research findings provide new insights for the AI field, enhancing the authenticity and reasoning capabilities of language models."]