Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
Segment Anything Model 2github.comSegment Anything Model (SAM) is a foundation model developed by Meta AI Research that produces high-quality object masks from input prompts like points or boxes. It can generate masks for all objects in an image and has strong zero-shot performance across various segmentation tasks.
Sources
geoaiunpacked.substack.com shapes more of what AI says about Segment Anything Model 2 than any other source, at 20% of its citations.
linkedin.com · roboflow.com · twinworks.ai · youtube.com
The market map
Data Annotation Platforms and Services →Where AI ranks Segment Anything Model 2
