Find edge AI hardware
Task keyword spotting · model-size class unknown · batch 1 · 17 platforms
Not like-for-like across groups: same model-size class and batch, but 5 different model/precision combinations. Compare within a group; each group is sorted by result.
DS-CNN (MLPerf Tiny KWS) · PTQ (2)
| Platform | Result | Model · precision · stack | Grade · date | Power | Price | Source |
|---|---|---|---|---|---|---|
| Next Generation Snapdragon Mobile Platform MTP | 15396 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) PTQ · Qualcomm AI Stack | A 2025-09-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.3 (closed) |
| Snapdragon 8 Gen 3 Mobile HDK (Snapdragon 8 Gen 3 Mobile Platform) | 11705 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) PTQ · Qualcomm AI Stack | A 2024-04-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.2 (closed) |
DS-CNN (MLPerf Tiny KWS) · Uniform symmetric quantization (2)
| Platform | Result | Model · precision · stack | Grade · date | Power | Price | Source |
|---|---|---|---|---|---|---|
| K1_BOARD_V1.0_HIGH_FREQ (K1 Octa-Core RISC-V) | 1735 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) Uniform symmetric quantization · | A 2025-09-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.3 (closed) |
| K1_BOARD_V1.0 (K1 Octa-Core RISC-V) | 717 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) Uniform symmetric quantization · | A 2025-09-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.3 (closed) |
DS-CNN (MLPerf Tiny KWS) · Uniform scaling from pre-trained weights (1)
| Platform | Result | Model · precision · stack | Grade · date | Power | Price | Source |
|---|---|---|---|---|---|---|
| NDP9120 (N/A) | 675 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) Uniform scaling from pre-trained weights · N/A | A 2025-09-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.3 (closed) |
DS-CNN (MLPerf Tiny KWS) · INT8 (7)
| Platform | Result | Model · precision · stack | Grade · date | Power | Price | Source |
|---|---|---|---|---|---|---|
| EK-RA8D1 (R7FA8D1AHECBD) | 289 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) INT8 · | A 2024-04-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.2 (closed) |
| NUCLEO-H7A3ZI-Q (STM32H7A3ZIT6Q) | 88.36 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) INT8 · X-CUBE-AI v9.0.0 | A 2024-04-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.2 (closed) |
| NUCLEO-U3C5ZIT6-Q | 60.49 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) INT8 · STEDGE-AI Internal version | A 2025-09-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.3 (closed) |
| NUCLEO-U575ZI-Q (STM32U575ZIT6Q) | 34.49 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) INT8 · X-CUBE-AI v10.2 | A 2025-09-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.3 (closed) |
| NUCLEO-L4R5ZI (STM32L4R5ZIT6U) | 23.16 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) INT8 · X-CUBE-AI v9.0 | A 2024-04-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.2 (closed) |
| NUCLEO-L4R5ZI-P (STM32L4R5ZIT6P) | 23.03 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) INT8 · X-CUBE-AI v10.2 | A 2025-09-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.3 (closed) |
| NUCLEO-U385RG-Q (STM32U385RGT6Q) | 20.70 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) INT8 · X-CUBE-AI v10.2 | A 2025-09-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.3 (closed) |
DS-CNN (MLPerf Tiny KWS) · not stated (5)
| Platform | Result | Model · precision · stack | Grade · date | Power | Price | Source |
|---|---|---|---|---|---|---|
| NUCLEO-H7A3ZI (STM32H7A3ZIT6Q) | 68.08 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) not stated · Bosch Hardware-Aware Lowering Engine (HALE) version 1.0.175 | A 2024-04-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.2 (closed) |
| DISCO-F746NG (STM32F746NG) | 52.75 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) not stated · Bosch Hardware-Aware Lowering Engine (HALE) version 1.0.175 | A 2024-04-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.2 (closed) |
| CY8CPROTO-062-4343W (PSoC 62 MCU) | 18.47 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) not stated · Bosch Hardware-Aware Lowering Engine (HALE) version 1.0.175 | A 2024-04-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.2 (closed) |
| Renesas RH850/F1KM-S4 R7F701649 | 15.71 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) not stated · Bosch Hardware-Aware Lowering Engine (HALE) version 1.0.175 | A 2024-04-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.2 (closed) |
| NUCLEO-G0B1RE (STM32G0B1RE) | 2.11 inferences/s throughput b=1 | DS-CNN (MLPerf Tiny KWS) not stated · Bosch Hardware-Aware Lowering Engine (HALE) version 1.0.175 | A 2024-04-17 (published) | unknown no power published | unknown | MLPerf Tiny v1.2 (closed) |
Confidence: A audited MLPerf closed, B documented third-party, C community, D vendor-published, E estimate. Unknown is never assumed. Other like-for-like groups: unknown, batch 1 (17)