Benchmark Explorer · Hailo-8

Hailo-8: 52 benchmark rows, 69% measured.

Method v1.4.1 · dataset 2026-09-29 · verified 2026-09-16 · last updated September 2026

36 of 52 rows are measured and published (class C); the rest are EdgeAIStack-derived (class D). 17 model families. Precisions: INT8 (52). Runtimes: hailo sdk (52). Tasks: detection (31), classification (4), pose (6), segmentation (11).

Family × precision: best value

The fastest matching row per family and precision (fps, or tokens/s for language models), with its evidence class. Click a cell for the full breakdown.

Family	INT8
CenterNet	405 fps C
DETR	30.5 fps C
EfficientDet	95 fps C
MobileNet	3305 fps C
SSD MobileNet	800 fps D
MobileNetV2	2597 fps C
ResNet	1372 fps C
SSD MobileNet	1015 fps C
YOLO11	541 fps C
YOLO11	167 fps D
YOLO11	130 fps D
YOLOv10	567 fps C
YOLOv5	543 fps C
YOLOv5	464 fps C
YOLOv8	1036 fps C
YOLOv8	943 fps D
YOLOv8	528 fps C

Hailo-8: 52 benchmark rows, 69% measured.

Family × precision: best value

Every model on Hailo-8

Links

Query this hardware live.