A text and image processing model that operates without content filtering, which is its defining characteristic. It handles both visual and textual inputs, producing text outputs. The absence of safety guardrails means it will engage with content that filtered models decline, though this also means users bear full responsibility for appropriate use.