Trending storyDeveloping

AgentScope releases Piccolo open-weight decision models in three sizes

1 article1 sourcesince Nov 5Last article 1h ago · Nov 5, 5:09 AM ET

Overview

AISummary of 1 article

AgentScope introduces Piccolo, a set of open-weight decision models in 27B, 9B and 4B sizes that return structured decisions with probabilities for agent and data workflows, according to a ModelScope post on X. The post says Piccolo-27B scores 92.21% on JevBench Public and 84.41% on HarnessJevBench for agent execution control. The models are released under Apache 2.0, and third-party licenses apply to any components that carry them.

Written by AI from the articles below · updated Nov 5, 5:16 AM ET

Check the sources:

Latest developmentNov 5, 5:09 AM ET

AgentScope releases Piccolo open-weight decision models in three sizes

Article timeline

The articles in this story. Times are ET.

Nov 5

5:09 AM
ModelScope@ModelScope2022Official
AgentScope releases Piccolo open-weight decision models in three sizes
AIAgentScope introduces Piccolo, open-weight decision models in 27B, 9B, and 4B sizes that return structured decisions with probabilities for agent and data workflows. Piccolo-27B scores 92.21% accuracy on JevBench Public and 84.41% on HarnessJevBench for agent execution control. The models are released under Apache 2.0, with third-party licenses applying.

Heat Trend

Not enough continuous observations to show a trend yet.