Model detail
Segment Anything Model 3
Meta's current public SAM 3 open-weight promptable segmentation model for images and video.
segmentationobject_trackingvisual_segmentationresearchlocal_inference
AI Model Radar / AI模型雷达
Segment Anything Model 3
Model detail, capability tags, source, OpenRouter state, and public reputation.
Source
facebook/sam3 Hugging Face model card API
P0 · official · high
Current version
SAM 3
CompanyMeta
ReleaseNot disclosed
ContextNot disclosed
Parameters0.86B
| Input modalities | image / video / text |
|---|---|
| Output modalities | structured_output |
| Tool use / Function calling | No reliable public source / No reliable public source |
| Structured output / Reasoning | Yes / No reliable public source |
| Vision / Audio / Video | Yes / No / Yes |
| Open weights / License | Yes / Not applicable |
| OpenRouter | No reliable public source |
| Community reputation | 1 public signals |
Versions