智通财经APP获悉，9月22日，美团(03690)发布高效推理模型LongCat-Flash-Thinking。美团表示，基于AIME25实测数据，LongCat-Flash-Thinking在该框架下展现出更高效的智能体工具调用能力，在确保90%准确率的前提下，相较于不使用工具调用节省了64.5%的Tokens。目前，该模型已在HuggingFace、Github全面开源。
官方介绍，该模型不仅增强了智能体自主调用工具的能力，还扩展了形式化定理证明能力，成为国内首个同时具备“深度思考+工具调用”与“非形式化+形式化”推理能力相结合的大语言模型。尤其在超高复杂度的任务(如数学、代码、智能体任务)处理上，LongCat-Flash-Thinking具备更显著的优势。
综合评估显示，LongCat-Flash-Thinking在逻辑、数学、代码、智能体等多个领域的推理任务中，达到了全球开源模型的最先进水平(SOTA)。
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.