Abu Dhabi-based edge AI startup Apertix has open-sourced Cerno 9B v1.2, a small decision model designed for agentic AI tasks running on edge devices.
The model ranks second among open models with fewer than 25 billion parameters on the Decision Index Full Score leaderboard, behind baselabs’ elephant-1. According to the Decision Index, Cerno 9B has a median latency of 74 milliseconds, compared with 114 milliseconds for elephant-1.
Rather than generating text like a large language model, Cerno reads a situation and returns probabilities for a fixed set of options in one pass. This allows it to support workflows such as routing requests, classifying inputs, selecting tools and deciding an AI agent’s next action.
Apertix said Cerno 9B scored 82.0 on tool use, ahead of Jev’s 74.4. Its public Decision Index score rose to 66.83 from 58.59 for the previous version.
The model’s weights and inference code are available on Hugging Face under an Apache 2.0 licence. Cerno is part of a broader family that also includes 27B and 4B models. Apertix is participating in the AI incubator at the MBZUAI Incubation and Entrepreneurship Center.
Source: Middle East AI News
