Artificial Analysis Decision Index
An evaluation metric introduced by Artificial Analysis for capability on decision-making tasks. It is designed to measure model performance in decision-making scenarios, providing a standardized basis for comparing how well different models handle tasks that require choosing among options or arriving at a decision.
Timeline (last 90 days)
Elephant-1 is a 9B-parameter model and is tied for first place with other models on the Artificial Analysis Decision Index.
The Jev Decision Index and the Artificial Analysis Decision Index are third-party leaderboards designed for decision-type tasks, and their evaluations cover dimensions such as accuracy and consistency in intent classification, entity-extraction-style judgments, and multi-label scoring, rather than generation quality or general knowledge question answering.
According to a MarkTechPost report, d1-ultra scored 65 on the Artificial Analysis Decision Index, with an input cost of $0.042 per million tokens.
The specific composition of the Artificial Analysis Decision Index's evaluation set, the distribution of task types, and the scoring methodology are only limitedly disclosed in public information.
All Facts (4)
Elephant-1 is a 9B-parameter model and is tied for first place with other models on the Artificial Analysis Decision Index.
50%UnverifiedThe Jev Decision Index and the Artificial Analysis Decision Index are third-party leaderboards designed for decision-type tasks, and their evaluations cover dimensions such as accuracy and consistency in intent classification, entity-extraction-style judgments, and multi-label scoring, rather than generation quality or general knowledge question answering.
50%UnverifiedAccording to a MarkTechPost report, d1-ultra scored 65 on the Artificial Analysis Decision Index, with an input cost of $0.042 per million tokens.
50%UnverifiedThe specific composition of the Artificial Analysis Decision Index's evaluation set, the distribution of task types, and the scoring methodology are only limitedly disclosed in public information.
50%