Perceptron Mk1

A vision-language model from Perceptron designed to process images and videos alongside natural language queries for visual understanding tasks.

Strengths

Handles multimodal inputs including video and still images, enabling detailed analysis of visual content and spatial reasoning across temporal sequences.

Best for

Applications requiring video analysis, embodied reasoning, or detailed visual understanding from image and video inputs paired with natural language instructions.

Limitations

Optimized for vision and embodied reasoning; may not match specialized text-only models on pure language tasks unrelated to visual content.