SAM 3
HoldPlatforms
A segmentation model used here as a comparison baseline.
Why it's here
Placed in Hold: 1 article(s) of evidence from 1 source(s), led by research-stage coverage, with 0 in the last 30 days. Confidence 24%. Low accumulated evidence, so it defaults conservatively pending more signal.
Evidence (1)
- 7Hugging Face Blog·4/1/2026researchFalcon Perception: Early-Fusion Transformer for Grounding and Segmentation
Falcon Perception is a 0.6B-parameter early-fusion Transformer for open-vocabulary grounding and segmentation from natural language prompts. Hugging Face reports 68.0 Macro-F1 on SA-Co, along with a new diagnostic benchmark called PBench, and also introduces Falcon OCR, a 0.3B-parameter OCR model with strong benchmark results and high throughput.