Meta AI Releases Segment Anything Model 3 (SAM 3) for Promptable Concept Segmentation in Images and Videos

Meta AI has introduced the open-source Meta Segment Anything Model 3 (SAM 3), a unified foundation model designed for promptable segmentation of images and videos that operates directly on visual concepts rather than just pixels. Unlike its predecessor SAM 2, SAM 3 can comprehensively identify, segment, and track all instances of an open vocabulary conceptsuch as every "red baseball cap" in a lengthy videousing a single model, thereby enabling scalable large-scale concept detection. The model advances the concept of Promptable Concept Segmentation (PCS), which accepts both text prompts and visual exemplarssuch