Geekbench AI Score

864
Single Precision Score
881
Half Precision Score
1744
Quantized Score
Geekbench AI 1.2.0 for Android AArch64

Result Information

Upload Date December 28 2024 12:32 PM
Views 1

System Information

AI Information
Framework TensorFlow Lite
Backend CPU
Device MediaTek Dimensity 7200 Pro
System Information
Operating System Android 15
Model Nothing Phone (2a)
Model IDNothing A142
Motherboard Pacman
CPU Information
Name ARM ARMv8
Topology 1 Processor, 8 Cores
Identifier ARM implementer 65 architecture 8 variant 1 part 3405 revision 0
Base Frequency 2.00 GHz
Cluster 16 Cores @ 2.00 GHz
Cluster 22 Cores @ 2.80 GHz
Memory Information
Size 11.16 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 697
129.7 IPS
 
Image Classification (HP)
100% 702
130.6 IPS
 
Image Classification (Q)
100% 1293
240.5 IPS
 
Image Segmentation (SP)
100% 660
10.7 IPS
 
Image Segmentation (HP)
100% 667
10.8 IPS
 
Image Segmentation (Q)
98% 1281
20.8 IPS
 
Pose Estimation (SP)
100% 1301
1.52 IPS
 
Pose Estimation (HP)
100% 1296
1.51 IPS
 
Pose Estimation (Q)
84% 3371
4.00 IPS
 
Object Detection (SP)
98% 637
50.7 IPS
 
Object Detection (HP)
98% 638
50.8 IPS
 
Object Detection (Q)
83% 1132
91.4 IPS
 
Face Detection (SP)
100% 1062
12.6 IPS
 
Face Detection (HP)
100% 1065
12.7 IPS
 
Face Detection (Q)
95% 2622
31.3 IPS
 
Depth Estimation (SP)
99% 963
7.44 IPS
 
Depth Estimation (HP)
99% 1197
9.25 IPS
 
Depth Estimation (Q)
64% 2701
24.7 IPS
 
Style Transfer (SP)
89% 3210
4.17 IPS
 
Style Transfer (HP)
89% 3208
4.16 IPS
 
Style Transfer (Q)
98% 7218
9.31 IPS
 
Image Super-Resolution (SP)
100% 584
21.6 IPS
 
Image Super-Resolution (HP)
100% 589
21.8 IPS
 
Image Super-Resolution (Q)
97% 1702
63.1 IPS
 
Text Classification (SP)
100% 375
500.3 IPS
 
Text Classification (HP)
99% 365
486.7 IPS
 
Text Classification (Q)
88% 969
1.31 KIPS
 
Machine Translation (SP)
100% 849
14.6 IPS
 
Machine Translation (HP)
100% 832
14.3 IPS
 
Machine Translation (Q)
50% 491
17.4 IPS
 

Compare

Set Baseline Geekbench 6 now available