Article

Ant Digital Technologies releases a decision-making model LingDT-Decision-1.0, with accuracy surpassing Microsoft and JEV and other models

PANews October 22 reports that Ant Digital Technologies has officially released its decision-making model LingDT-Decision-1.0, offering 4B and 27B versions. It is now available on Ant Digital Technologies' DTMaaS platform, where developers can try it free for a limited time. Designed specifically for rapid decision-making tasks, the model shifts large models from a "generate first, then parse" approach to "making direct judgments among constrained options." In real-world business tests across customer service, risk control, finance, and other fields,Token consumption was reduced by more than 90%, while response efficiency improved by up to 10 times. In intelligent customer service scenarios, risk-identification accuracy increased from 76%to 89%, while hallucination recall in financial scenarios increased from 44%to 70%。LingDT-Decision-1.0 also claims to be the world's first to offer native multi-select and sequential decision-making (DAG) capabilities, enabling it to identify multiple requests at once and complete multi-step decisions according to task dependencies, making it better suited to...