Institute of Foundation Models releases K2 Horizon, six fully open models from 0.9B to 375B

The Institute of Foundation Models, based in Abu Dhabi, introduced K2 Horizon on Wednesday, a fleet of six AI foundation models ranging from 0.9 billion to 375 billion parameters. The institute says the models are fully open, including weights, code, training data and methodologies, so researchers and developers can inspect, reproduce and adapt them. The lineup includes a 0.9B model designed for highly constrained hardware such as watches and glasses, 3.7B and 7B models intended for phones and other on-device uses, a dense 32B model and a sparse 36B-A4B model for local hosting and on-premise servers, and a 375B-A23B model for demanding enterprise deployments. The institute claims top-tier performance in every size class across reasoning, mathematics, coding and agentic tasks, and says the 0.9B, 3.7B and 7B models set a new state of the art at their respective scales. Those performance claims come from the institute's own announcement and have not been independently verified in the available reporting. The release is notable for its scope of openness, since many open-weight models withhold training data or training recipes, which limits reproducibility. Full disclosure of data and methods gives academic and industry researchers a basis for auditing and extending the models.