According to CoinMeta, the Sage2 developed by the American startup Levanto Labs allows for simultaneous chatting and updating of judgments, with a fastest response time of 50 milliseconds. This system combines information extraction with real-time decision-making, capable of automatically extracting details such as amounts, dates, and user needs from conversations, and continuously adjusting its judgments based on new content. For example, customer service representatives on websites can use Sage2 to monitor users' emotions in real-time; when user dissatisfaction reaches a set threshold, the system can automatically transfer the call to a human representative. Officials claim that Sage2 takes about 150 milliseconds to extract information from short texts, making it suitable for handling a large number of repetitive judgment tasks. Compared to general-purpose large models, specialized decision-making models are expected to reduce latency and costs. The information extraction accuracy score of Sage2 is 77.8, slightly higher than the 75.3 of Clairic, the data extraction model of Cloudflare.