Apple releases Depth Pro 2 for sharp, temporally consistent depth estimation
, updated First seen Updated Broad coverage, tapering
Apple · ModelsAAPL$342.56▼1.2%
Reported by Apple Machine Learning Research, explainx.ai, Digg and 1 more outlets4 sources ↓
Apple says Depth Pro 2 estimates metric depth from images and video at native resolution, combining sharp boundaries with temporal consistency. The company says it is releasing a family of models.
Key points
- The models process both images and video and predict depth maps at native resolution
- Apple says the models combine metric accuracy, sharpness and temporal consistency
- Apple says it is releasing multiple models with different runtime and accuracy trade-offs
Sources
- Apple Machine Learning ResearchRead ↗Depth Pro 2: Sharp and Temporally Consistent Metric Monocular Depth
- explainx.aiRead ↗Apple Depth Pro 2: 4 Open Depth Models, 24M to 872M
- DiggRead ↗Apple releases Depth Pro 2—and the weights aren’t for commercial apps
- OrcaRouterRead ↗Depth Pro 2 Shipped Without a Paper: What the Weights Show
Details
Apple Machine Learning Research describes Depth Pro 2 as a network for estimating dense metric depth from images and video. The lab says earlier methods often traded off temporal consistency against sharp boundaries, and says Depth Pro 2 is designed to provide both.
The article does not provide benchmark scores, model availability details or licensing terms.
Written from coverage as of
Images
- Apple Machine Learning Research
- Apple Machine Learning Research
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh