Apertus v1.5 70B FP8 is a large, open-weight multimodal model from onprem-ai that accepts both text and image inputs. Running in FP8 precision, it trades a small amount of numerical fidelity for reduced memory footprint, making a 70B-scale model more practical to deploy on local hardware. Specific capability details are limited, but its scale and multimodal inputs suggest broad general-purpose use.