depth estimation benchmarks on edge AI hardware
73 published rows. Each table compares only results with the same model-size class and batch size. Search this workload · JSON.
Fastest in the best-covered like-for-like group (nano (YOLO-n), batch 1): Apple M4 (10-core GPU, 120 GB/s) at 133 FPS B 2026-10-01 (page last edited). 8 groups, 12 platforms.
Show data — 8 like-for-like tables
nano (YOLO-n) · batch 1 (5 platforms)
✓ Like-for-like: model-size class nano (YOLO-n), batch 1. Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| Apple M4 (10-core GPU, 120 GB/s) | 133 FPS latency_e2e · b=1 | YOLO26n-depth FP16 · Ultralytics 8.4.168 Core AI (coreai-torch 0.4.3), macOS 27.0 · nano (YOLO-n) | B | unknown power n/a no power published | Ultralytics documentation: Apple Core AI integration benchmark 2026-10-01 (page last edited) |
| Axelera Metis Embedded 113m M.2 card | 96.00 FPS throughput · b=1 | YOLO26n-Depth INT8 · Voyager SDK (GStreamer pipeline) · nano (YOLO-n) | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
| Axelera Metis Edge 130p PCIe card | 90.00 FPS throughput · b=1 | YOLO26n-Depth INT8 · Voyager SDK (GStreamer pipeline) · nano (YOLO-n) | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
| Raspberry Pi 5 + AI HAT+ (Hailo-8) | 33.78 FPS inference_time · b=1 | YOLO26n-depth INT8 · Ultralytics Hailo HEF / HailoRT · nano (YOLO-n) | B | unknown power n/a no power published | Ultralytics documentation: Hailo integration benchmark 2026-10-11 (page last edited) |
| Raspberry Pi 5 (CPU only) CPU-only | 0.470 FPS inference_time · b=1 | YOLO26n-depth FP32 · Ultralytics PyTorch (CPU) · nano (YOLO-n) | B | unknown power n/a no power published | Ultralytics documentation: Hailo integration benchmark 2026-10-11 (page last edited) |
small (YOLO-s) · batch 1 (2 platforms)
✓ Like-for-like: model-size class small (YOLO-s), batch 1. Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| Axelera Metis Embedded 113m M.2 card | 77.00 FPS throughput · b=1 | YOLO26s-Depth INT8 · Voyager SDK (GStreamer pipeline) · small (YOLO-s) | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
| Axelera Metis Edge 130p PCIe card | 71.00 FPS throughput · b=1 | YOLO26s-Depth INT8 · Voyager SDK (GStreamer pipeline) · small (YOLO-s) | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
medium (YOLO-m) · batch 1 (2 platforms)
✓ Like-for-like: model-size class medium (YOLO-m), batch 1. Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| Axelera Metis Embedded 113m M.2 card | 46.00 FPS throughput · b=1 | YOLO26m-Depth INT8 · Voyager SDK (GStreamer pipeline) · medium (YOLO-m) | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
| Axelera Metis Edge 130p PCIe card | 44.00 FPS throughput · b=1 | YOLO26m-Depth INT8 · Voyager SDK (GStreamer pipeline) · medium (YOLO-m) | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
large (YOLO-l) · batch 1 (2 platforms)
✓ Like-for-like: model-size class large (YOLO-l), batch 1. Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| Axelera Metis Embedded 113m M.2 card | 36.00 FPS throughput · b=1 | YOLO26l-Depth INT8 · Voyager SDK (GStreamer pipeline) · large (YOLO-l) | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
| Axelera Metis Edge 130p PCIe card | 33.00 FPS throughput · b=1 | YOLO26l-Depth INT8 · Voyager SDK (GStreamer pipeline) · large (YOLO-l) | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
xlarge (YOLO-x) · batch 1 (2 platforms)
✓ Like-for-like: model-size class xlarge (YOLO-x), batch 1. Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| Axelera Metis Embedded 113m M.2 card | 20.00 FPS throughput · b=1 | YOLO26x-Depth INT8 · Voyager SDK (GStreamer pipeline) · xlarge (YOLO-x) | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
| Axelera Metis Edge 130p PCIe card | 19.00 FPS throughput · b=1 | YOLO26x-Depth INT8 · Voyager SDK (GStreamer pipeline) · xlarge (YOLO-x) | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
small ViT · batch 1 (1 platforms)
✓ Like-for-like: model-size class small ViT, batch 1. Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| Axelera Metis Edge 130p PCIe card | 20.73 FPS inference_time · b=1 | dinov2_depth_vits14_nyu_linear not stated · Voyager SDK model_axscript · small ViT | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
unknown · batch 1 (4 platforms)
✓ Like-for-like: model-size class unknown, batch 1 (size or batch not stated by source — unverified). Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| Axelera Metis Edge 130p PCIe card | 946 FPS throughput · b=1 | FastDepth INT8 · Voyager SDK (GStreamer pipeline) · unknown | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
| Axelera Metis Embedded 113m M.2 card | 942 FPS throughput · b=1 | FastDepth INT8 · Voyager SDK (GStreamer pipeline) · unknown | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
| MemryX MX3 M.2 module | 501 FPS throughput · b=1 | PyDNet not stated (MX3 activations BF16 / weights 4-8-16 bit per vendor) · MemryX SDK 2.2 (closed-loop compiled DFP) · unknown | C | unknown power n/a no power published | MemryX Model Explorer (SDK 2.2) 2026-07-15 (published) |
| Microchip PolarFire SoC Video Kit (VectorBlox CoreVectorBlox) | 13.39 FPS inference_time · b=1 | Midas-V2_256x128 INT8 (VNNX) · VectorBlox SDK (VNNX), no compression · unknown | D | unknown power n/a no power published | Microchip VectorBlox SDK tutorial metrics 2026-09-11 (page last edited) |
unknown · batch unknown (5 platforms)
✓ Like-for-like: model-size class unknown, batch unknown (size or batch not stated by source — unverified). Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| AMD Versal AI Edge VEK280 (VE2802) | 19.15 FPS E2E throughput, multi thread · b=unknown | pt_fadnetv2_3.5 INT8 · Vitis AI 3.5 (VART) · unknown | D | unknown power n/a no power published | Vitis AI 3.5 Model Zoo performance table Older data (2023) |
| AMD Versal AI Core VCK190 (VC1902) | 18.73 FPS E2E throughput, multi thread · b=unknown | pt_fadnetv2_sceneflow_576_960_0.51_201G INT8 · Vitis AI 3.0 (VART) · unknown | D | unknown power n/a no power published | Vitis AI 3.0 Model Zoo performance table Older data (2023) |
| AMD Zynq UltraScale+ ZCU104 (ZU7EV) | 7.89 FPS E2E throughput, multi thread · b=unknown | pt_fadnetv2_sceneflow_576_960_0.51_201G INT8 · Vitis AI 3.0 (VART) · unknown | D | unknown power n/a no power published | Vitis AI 3.0 Model Zoo performance table Older data (2023) |
| AMD Zynq UltraScale+ ZCU102 (ZU9EG) | 4.57 FPS E2E throughput, multi thread · b=unknown | pt_fadnetv2_sceneflow_576_960_0.51_201G INT8 · Vitis AI 3.0 (VART) · unknown | D | unknown power n/a no power published | Vitis AI 3.0 Model Zoo performance table Older data (2023) |
| AMD Kria KV260 Vision AI Starter Kit (K26 SOM) | 4.51 FPS E2E throughput, multi thread · b=unknown | pt_fadnetv2_sceneflow_576_960_0.51_201G INT8 · Vitis AI 3.0 (VART) · unknown | D | 7.50 W spec reported in result source — spec | Vitis AI 3.0 Model Zoo performance table Older data (2023) |
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.