美团上线首个开源并可体验的“重思考”模型，工具调用能力登顶开源 SOTA

Core Insights - The LongCat-Flash-Thinking-2601 model has been released as an upgraded version of the LongCat-Flash-Thinking model, now available as open source [1][4] - This new model achieves state-of-the-art (SOTA) performance on key evaluation benchmarks related to agent search, tool invocation, and tool interaction reasoning [1][4] - The model outperforms Claude-Opus-4.5-Thinking in random complex tasks that rely on tool invocation, significantly reducing the training costs for adapting new tools in real-world scenarios [1][4] - The model features a "rethink" mode that allows for the simultaneous activation of 8 "brains" to execute tasks, which can be experienced for free on the LongCat official website [1][4]