Ant Digital Technologies today released the LingDT-Decision-1.0 decision model. Designed specifically for rapid industry decision-making, this model offers higher accuracy, faster response times, and lower token costs when handling structured decision tasks.

LingDT-Decision-1.0 is the first decision model to natively support multi-select and sequential decision-making capabilities, making it better suited for real-world enterprise applications. Across 23 open-source evaluation benchmarks, LingDT-Decision-1.0 outperforms popular decision models such as Microsoft-Decision-1 and Jev 1.13, with its 27B version achieving the highest accuracy among models of comparable parameter size.
Testing by the Ant Digital Science team in real business scenarios such as customer service, risk control, and finance shows that, when handling the same tasks, LingDT-Decision-1.0 reduces token consumption by over 90%, improves response speed by 10 times, and delivers superior performance compared to large language models.
In real-world enterprise applications, there are numerous high-frequency decision-making tasks. For example, in a customer service scenario, when receiving the inquiry, “This order hasn’t been shipped after three days—please handle it urgently,” the AI must quickly make a series of judgments and decisions: categorizing the issue type (refund, order follow-up, or complaint?), assessing ticket urgency (does it require priority handling?), assigning the appropriate department (which team should handle the ticket?), and evaluating risk (does it require human intervention?).
Traditional large language models have significant limitations in such scenarios. They typically require token-by-token reasoning before generating output text, resulting in long pipelines, high costs, increased error rates, and unstable performance. Decision models are specifically designed to address these pain points—their goal is not to generate text, but to make fast, accurate decisions. As a result, they are naturally suited for high-frequency decision-making tasks in enterprise business, such as intent recognition, tool selection, rule classification, and scoring outcomes.
Compared to other decision models, Ant Digital Technologies LingDT-Decision-1.0 places greater emphasis on being designed for real-world enterprise applications. In addition to foundational capabilities such as judgment, single-choice selection, classification, and rating, LingDT-Decision-1.0 introduces, for the first time, native multi-choice and sequential decision-making (DAG) functions, better accommodating complex, multi-step decision-making needs in enterprise business processes.

LingDT-Decision-1.0 is available in two versions: 4B and 27B. The model has now been launched on the Ant Digital Science DTMaaS platform, and developers can experience it for free for a limited time.
