Geekbench ML Score

349
TensorFlow Lite NNAPI Inference Score
Geekbench ML 0.6.0 for Android AArch64

Result Information

Upload Date October 14 2025 02:26 PM
Views 8

System Information

System Information
Operating System Android 15
Model Xiaomi 24122RKC7C
Model IDXiaomi 24122RKC7C
Motherboard sun
Governorwalt
CPU Information
Name Qualcomm ARMv8
Topology 1 Processor, 8 Cores
Identifier ARM implementer 81 architecture 8 variant 3 part 1 revision 4
Base Frequency 3.53 GHz
Cluster 16 Cores @ 3.53 GHz
Cluster 22 Cores @ 4.32 GHz
Memory Information
Size 14.76 GB
Inference Information
Framework TensorFlow Lite
Backend NNAPI
Device Default

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 134
25.1 IPS
 
Image Classification (F16)
100% 134
25.0 IPS
 
Image Classification (I8)
97% 360
67.3 IPS
 
Image Segmentation (F32)
100% 172
2.87 IPS
 
Image Segmentation (F16)
100% 172
2.87 IPS
 
Image Segmentation (I8)
98% 363
6.06 IPS
 
Pose Estimation (F32)
100% 239
0.29 IPS
 
Pose Estimation (F16)
100% 239
0.29 IPS
 
Pose Estimation (I8)
100% 906
1.10 IPS
 
Object Detection (F32)
100% 141
10.5 IPS
 
Object Detection (F16)
100% 142
10.6 IPS
 
Object Detection (I8)
61% 401
30.0 IPS
 
Face Detection (F32)
100% 378
4.49 IPS
 
Face Detection (F16)
100% 377
4.48 IPS
 
Face Detection (I8)
86% 818
9.73 IPS
 
Depth Estimation (F32)
100% 316
2.45 IPS
 
Depth Estimation (F16)
100% 317
2.45 IPS
 
Depth Estimation (I8)
95% 884
6.85 IPS
 
Style Transfer (F32)
100% 588
0.77 IPS
 
Style Transfer (F16)
100% 589
0.77 IPS
 
Style Transfer (I8)
98% 1627
2.14 IPS
 
Image Super-Resolution (F32)
100% 164
5.87 IPS
 
Image Super-Resolution (F16)
100% 164
5.87 IPS
 
Image Super-Resolution (I8)
98% 573
20.5 IPS
 
Text Classification (F32)
100% 341
490.6 IPS
 
Text Classification (F16)
100% 187
268.2 IPS
 
Text Classification (I8)
92% 526
755.8 IPS
 
Machine Translation (F32)
100% 616
11.3 IPS
 
Machine Translation (F16)
100% 621
11.4 IPS
 
Machine Translation (I8)
64% 628
11.6 IPS
 

Compare

Set Baseline Geekbench 6 now available