On September 22, Meituan (03690) released the efficient inference model LongCat-Flash-Thinking. Meituan stated that based on AIME25 real-world data, LongCat-Flash-Thinking demonstrates a more efficien ...