Experimental research model with enhanced visual reasoning capabilities.

Supports context length of 128k.

Currently, the model only supports single-round dialogues and image outputs. It does not support video inputs.

Should be capable of images up to 12 MP.