Meta announced the release of Segment Anything Model 3 (SAM 3.1), an updated version of its computer vision model designed to enhance video processing efficiency. The new model introduces object multiplexing, enabling the tracking of up to 16 objects in a single forward pass. This advancement doubles the processing speed for videos with a medium number of objects, increasing throughput from 16 to 32 frames per second on a single H100 GPU. As a result, SAM 3.1 supports real-time object tracking in complex videos while reducing GPU resource requirements, making high-performance applications more accessible on smaller hardware. The improvement stems from a change in how the model handles multiple objects, previously requiring dedicated passes for each object. With multiplexing, the model processes all tracked objects together, eliminating redundant computation and memory bottlenecks. This global reasoning approach streamlines performance and enhances accuracy in crowded scenes. Meta encourages the community to download the SAM 3.1 model checkpoint, explore the updates to the SAM 3 codebase and research paper, and test the model on the Segment Anything Playground. *Source: [metaai](https://ai.meta.com/blog/segment-anything-model-3/)*