Recently, Zhipu Technology introduced a groundbreaking AI search tool based on chain-of-thought, which has garnered significant attention within the industry due to its robust capabilities. This tool not only integrates deep web reading and multi-step reasoning but also seamlessly connects with other Zhipu AI tools, offering users an unprecedented search experience.
Following the release of OpenAI's O1 and O1mini models, the value of chain-of-thought and model reasoning capabilities has been widely recognized. These technologies have shown great potential in fact-checking and solving complex problems. Among various applications, search is undoubtedly one of the closest needs to ordinary users.
Zhipu's newly launched AI search tool aligns with this trend by combining reasoning capabilities with long-text reading, effectively addressing the common "hallucination" issues in previous AI searches. More importantly, it can summarize search results in a complex manner, providing users with deeper and more accurate information, significantly saving time costs for sorting and logical reasoning.
Key features of this AI search tool include:
Capable of reading over 100 web pages in one go, significantly enhancing information acquisition.
Supports multi-level reasoning chain-of-thought, offering more comprehensive solutions to complex problems.
Seamless integration with other Zhipu AI tools, enriching functionality.
In practical applications, the tool has demonstrated astonishing capabilities. For instance, users can add calculation-based questions on complex content, such as asking, "What percentage of a normal human's daily calorie intake is consumed by eating three passion fruits?" The tool not only provides accurate answers but also details the calculation process.
Even more impressive is its reasoning capability. Users can ask multiple logically related questions in one query, such as "Who invented the chip, what were their contributions, and what impact did it have on AI?" The AI search tool can comprehensively answer these interconnected questions, showcasing strong information integration and analytical skills.
Beyond information organization, the tool can also process content deeply. For example, when asked to analyze historical bull markets in the A-share market, it can not only read web information deeply and provide corresponding tables but also use Python to draw relevant charts, offering users intuitive data visualization results.
In answering highly specialized questions, such as "Who is the 'AI领域的麦橘' (the 'Wheat Orange' in the AI field)," Zhipu's AI search tool performs exceptionally well. It can accurately identify the essence of the question and provide detailed and accurate introductions, avoiding common misunderstandings and confusions.
