nvidia

NVIDIA Jetson Orin Nano

8 GB Orin SoM with 1024-core Ampere GPU + 32 Tensor Cores. The new robotics Pi 5.

top models on this chip

Object detection

Sorted by latency p50 · matched-pair runs on NVIDIA Jetson Orin Nano

Object detection

#	Model	Quant	Latency p50	Throughput	Acc.	Power
1	NVIDIA-YOLO27-Orin	INT8	15.5 ms	65 FPS	98.9%	13.1 W
2	RT-DETR-Edge	INT8	17.3 ms	58 FPS	98.2%	11.4 W
3	YOLO27-Edge	INT8	18.1 ms	55 FPS	97.9%	12.2 W
4	YOLO-World-Edge	INT8	22.6 ms	44 FPS	98.5%	13.1 W

top models on this chip

Depth

Cross-chip leaderboard →

Sorted by latency p50 · matched-pair runs on NVIDIA Jetson Orin Nano

Depth

#	Model	Quant	Latency p50	Throughput	Acc.	Power
1	MiDaS-Distilled-S	INT8	12.8 ms	78 FPS	98.8%	14.1 W
2	ZoeDepth-Mobile	INT8	18.3 ms	55 FPS	98.4%	12.1 W
3	Meridian Depth-Tier1	INT8	20.1 ms	50 FPS	98.3%	12.1 W
4	Depth-Anything-Edge	INT8	25.1 ms	40 FPS	97.3%	12.2 W

top models on this chip

Segmentation

Cross-chip leaderboard →

Sorted by latency p50 · matched-pair runs on NVIDIA Jetson Orin Nano

Segmentation

#	Model	Quant	Latency p50	Throughput	Acc.	Power
1	FastSAM-Edge	INT8	10.5 ms	95 FPS	99.0%	14.4 W

top models on this chip

Embedded ASR

Cross-chip leaderboard →

Sorted by latency p50 · matched-pair runs on NVIDIA Jetson Orin Nano

Embedded ASR

#	Model	Quant	Latency p50	Throughput	Acc.	Power
1	Distil-Whisper-Tiny	INT8	255.4 ms	25 tok/s	97.6%	12.9 W

top models on this chip

Multimodal

Cross-chip leaderboard →

Sorted by latency p50 · matched-pair runs on NVIDIA Jetson Orin Nano

Multimodal

#	Model	Quant	Latency p50	Throughput	Acc.	Power
1	MobileCLIP-S2	INT8	21.4 ms	47 FPS	97.8%	12.2 W
2	DINO-v3-Distilled	INT8	63.6 ms	16 FPS	97.4%	12.6 W
3	Florence-2-Edge	INT8	65.4 ms	15 FPS	98.2%	14.0 W
4	MobileCLIP-B	FP16	93.3 ms	11 FPS	99.4%	14.1 W
5	SmolVLM-3B	INT8	422.1 ms	2.4 FPS	98.7%	13.3 W

top models on this chip

Tracking

Cross-chip leaderboard →

Sorted by latency p50 · matched-pair runs on NVIDIA Jetson Orin Nano

Tracking

#	Model	Quant	Latency p50	Throughput	Acc.	Power
1	ByteTrack-Edge	INT8	6.16 ms	162 FPS	98.2%	12.3 W

top models on this chip

Pose

Cross-chip leaderboard →

Sorted by latency p50 · matched-pair runs on NVIDIA Jetson Orin Nano

Pose

#	Model	Quant	Latency p50	Throughput	Acc.	Power
1	RTMPose-Edge	INT8	9.95 ms	100 FPS	98.9%	14.2 W

top models on this chip

OCR

Cross-chip leaderboard →

Sorted by latency p50 · matched-pair runs on NVIDIA Jetson Orin Nano

OCR

#	Model	Quant	Latency p50	Throughput	Acc.	Power
1	DocLayout-Edge	INT8	13.8 ms	73 FPS	97.6%	12.0 W