Pre-trained Vision Model Products

YOLO11

Ultralytics

Ultralytics' flagship real-time detector family: five sizes (n→x) covering detection, instance segmentation, pose, OBB and classification, up to 54.7 mAP on COCO with ~11 ms TensorRT latency — the default choice for fast industrial object detection.

Grounding DINO

IDEA Research

Open-set object detector that finds objects from free-text prompts — no fine-tuning needed. 52.5 AP zero-shot on COCO; pairs with SAM 2 for text-prompted segmentation, the standard combo for flexible robotic perception.

SAM 2

Meta AI

Meta's foundation model for promptable segmentation in images and video — point/box prompts, multi-object tracking, streaming memory for real-time video, Apache-2.0.

FoundationPose

NVIDIA

NVIDIA's unified 6D object pose estimation and tracking model — model-based (CAD) or model-free (reference images), no fine-tuning for novel objects. #1 on the BOP leaderboard at release. Research use only.