HWBench

AMD Ryzen AI processor (SKU not stated) vs Hailo-8L (M.2/PCIe, x86 host)

Hailo-8L M.2 module from the Raspberry Pi AI Kit, mounted on the Raspberry Pi M.2 HAT+
Hailo-8L M.2 AI accelerator module (from the Raspberry Pi AI Kit). Photo: SparkFun Electronics, CC BY 2.0, via Wikimedia Commons.

Hailo-8L (M.2/PCIe, x86 host) is faster in all 2 like-for-like comparisons (largest gap: object detection, 3.13×).

AMD Ryzen AI processor (SKU not stated)Hailo-8L (M.2/PCIe, x86 host)
Best object detection144 FPS D
2025-12-17 (page last edited)		437 FPS D
2026-10-15 (page last edited)
Powern/a1.50 W spec
Perf / Wn/a291 FPS/W spec
Pricen/an/a
Workloads with data1 of 76 of 7

Best published object detection result per platform (may be from different like-for-like groups — see data). n/a = not published; never estimated.

Show data — all workloads, sources and like-for-like groups

object detection

✓ Like-for-like: model-size class small (YOLO-s), batch 1. Model, precision and stack may still differ.
PlatformResultModel · precision · stackGradePowerSource
Hailo-8L (M.2/PCIe, x86 host)137 FPS
throughput · b=1		yolov5s_wo_spp
INT8 (Hailo quantized) · Hailo DFC v2.20.0 / HailoRT (model zoo) · small (YOLO-s)		D1.50 W spec
reported in result source — spec		Hailo Model Zoo public models (Hailo-8L)
2026-10-15 (page last edited)
AMD Ryzen AI processor (SKU not stated)43.76 FPS
latency_e2e · b=1		YOLOv8s-worldv2
A16W8 · ONNX Runtime + Vitis AI EP (Ryzen AI SW), Quark A16W8 AdaRound · small (YOLO-s)		Dunknown power n/a
no power published		AMD technical article: Deploying object detection model on AMD AI PC
2026-01-23 (page last edited)
✓ Like-for-like: model-size class medium (YOLO-m), batch 1. Model, precision and stack may still differ.
PlatformResultModel · precision · stackGradePowerSource
Hailo-8L (M.2/PCIe, x86 host)62.90 FPS
throughput · b=1		yolov5m_wo_spp
INT8 (Hailo quantized) · Hailo DFC v2.20.0 / HailoRT (model zoo) · medium (YOLO-m)		D1.50 W spec
reported in result source — spec		Hailo Model Zoo public models (Hailo-8L)
2026-10-15 (page last edited)
AMD Ryzen AI processor (SKU not stated)48.20 FPS
throughput · b=1		yolov8m
XINT8 · Ryzen AI SW (ONNX Runtime, Vitis AI EP) run_inference.py --benchmark · medium (YOLO-m)		Dunknown power n/a
no power published		AMD RyzenAI-SW YOLOv8m example (README benchmark table)
2026-08-11 (published)

instance segmentation

Only Hailo-8L (M.2/PCIe, x86 host) has published instance segmentation results.

PlatformResultModel · precision · stackGradePowerSource
Hailo-8L (M.2/PCIe, x86 host)52.90 FPS
throughput · b=1		yolov5m_seg
INT8 (Hailo quantized) · Hailo DFC v2.20.0 / HailoRT (model zoo) · medium (YOLO-m)		D1.50 W spec
reported in result source — spec		Hailo Model Zoo public models (Hailo-8L)
2026-10-15 (page last edited)

AMD Ryzen AI processor (SKU not stated):

◌

No published instance segmentation results for AMD Ryzen AI processor (SKU not stated) or Hailo-8L (M.2/PCIe, x86 host) yet.
We only show cited public figures — nothing is estimated to fill gaps. Requests tell us what to source or test next.

pose estimation

Only Hailo-8L (M.2/PCIe, x86 host) has published pose estimation results.

PlatformResultModel · precision · stackGradePowerSource
Hailo-8L (M.2/PCIe, x86 host)795 FPS
throughput · b=1		mspn_regnetx_800mf
INT8 (Hailo quantized) · Hailo DFC v2.20.0 / HailoRT (model zoo) · mid-size CNN		D1.50 W spec
reported in result source — spec		Hailo Model Zoo public models (Hailo-8L)
2026-10-15 (page last edited)

AMD Ryzen AI processor (SKU not stated):

◌

No published pose estimation results for AMD Ryzen AI processor (SKU not stated) or Hailo-8L (M.2/PCIe, x86 host) yet.
We only show cited public figures — nothing is estimated to fill gaps. Requests tell us what to source or test next.

image classification

Only Hailo-8L (M.2/PCIe, x86 host) has published image classification results.

PlatformResultModel · precision · stackGradePowerSource
Hailo-8L (M.2/PCIe, x86 host)1738 FPS
throughput · b=1		mobilenet_v2_1.0
INT8 (Hailo quantized) · Hailo DFC v2.20.0 / HailoRT (model zoo) · mobile-class CNN		D1.50 W spec
reported in result source — spec		Hailo Model Zoo public models (Hailo-8L)
2026-10-15 (page last edited)

AMD Ryzen AI processor (SKU not stated):

◌

No published image classification results for AMD Ryzen AI processor (SKU not stated) or Hailo-8L (M.2/PCIe, x86 host) yet.
We only show cited public figures — nothing is estimated to fill gaps. Requests tell us what to source or test next.

semantic segmentation

Only Hailo-8L (M.2/PCIe, x86 host) has published semantic segmentation results.

PlatformResultModel · precision · stackGradePowerSource
Hailo-8L (M.2/PCIe, x86 host)19.60 FPS
throughput · b=1		fcn8_resnet_v1_18
INT8 (Hailo quantized) · Hailo DFC v2.20.0 / HailoRT (model zoo) · mid-size CNN		D1.50 W spec
reported in result source — spec		Hailo Model Zoo public models (Hailo-8L)
2026-10-15 (page last edited)

AMD Ryzen AI processor (SKU not stated):

◌

No published semantic segmentation results for AMD Ryzen AI processor (SKU not stated) or Hailo-8L (M.2/PCIe, x86 host) yet.
We only show cited public figures — nothing is estimated to fill gaps. Requests tell us what to source or test next.

speech-to-text

Only Hailo-8L (M.2/PCIe, x86 host) has published speech-to-text results.

PlatformResultModel · precision · stackGradePowerSource
Hailo-8L (M.2/PCIe, x86 host)41.80 FPS
throughput · b=1		whisper_base_5s_no_kqs_decoder
INT8 (Hailo quantized) · Hailo DFC v2.20.0 / HailoRT (model zoo) · Whisper base		D1.50 W spec
reported in result source — spec		Hailo Model Zoo public models (Hailo-8L)
2026-10-15 (page last edited)

AMD Ryzen AI processor (SKU not stated):

◌

No published speech-to-text results for AMD Ryzen AI processor (SKU not stated) or Hailo-8L (M.2/PCIe, x86 host) yet.
We only show cited public figures — nothing is estimated to fill gaps. Requests tell us what to source or test next.

LLM generation

◌

No published LLM generation results for AMD Ryzen AI processor (SKU not stated) or Hailo-8L (M.2/PCIe, x86 host) yet.
We only show cited public figures — nothing is estimated to fill gaps. Requests tell us what to source or test next.

How to read these numbers
Every number is labelled. Confidence: A MLPerf closed, audited · B documented third-party method · C community / review · D vendor-published, not independently verified · E estimate. Power basis: measured with the result · spec vendor spec/envelope · power n/a not published.

Machine-readable: /find.json (see llms.txt); full tables are in “Show data”. Confidence grades: A audited MLPerf closed · B documented third-party · C community · D vendor-published · E estimate. Unknown values are shown as unknown, never estimated. Generated 2026-10-16 from public sources.