SAM 3.1: Segment Anything Model

SAM 3.1 is Meta's latest Segment Anything Model with multiplex decoding, providing promptable image segmentation for ComfyUI.

SAM 3.1 (Segment Anything Model 3.1) is Meta's latest image segmentation model with multiplex decoding capabilities. It can segment any object in an image based on points, boxes, or text prompts. Repackaged for ComfyUI for use in segmentation and masking workflows.

Usage in ComfyUI

Place the model file in ComfyUI/models/checkpoints/. Requires a ComfyUI custom node that supports SAM 3.1 for loading and inference.

DeveloperMeta (Facebook AI Research)
Release Date2025
ArchitectureHierarchical Vision Transformer (Hiera Large) + Multiplex Decoding
LicenseSAM-License
TypeImage Segmentation
CapabilitiesPromptable segmentation (points, boxes, text)

Guides and workflows related to this model series.

No articles found.