Geekbench AI Score

603
Single Precision Score
603
Half Precision Score
1217
Quantized Score
Geekbench AI 1.6.1 for Android AArch64

Result Information

Upload Date January 26 2026 08:53 AM
Views 4

System Information

AI Information
Framework TensorFlow Lite
Backend CPU
Device ARM MT6877V/TZA
System Information
Operating System Android 13
Model Xiaomi 21091116UG
Model IDXiaomi 21091116UG
Motherboard pissarropro
CPU Information
Name ARM MT6877V/TZA
Topology 1 Processor, 8 Cores
Identifier ARM implementer 65 architecture 8 variant 1 part 3393 revision 0
Base Frequency 2.00 GHz
Cluster 16 Cores @ 2.00 GHz
Cluster 22 Cores @ 2.50 GHz
Memory Information
Size 7.31 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 282
52.4 IPS
 
Image Classification (HP)
100% 277
51.4 IPS
 
Image Classification (Q)
100% 834
155.1 IPS
 
Image Segmentation (SP)
100% 350
5.67 IPS
 
Image Segmentation (HP)
100% 301
4.88 IPS
 
Image Segmentation (Q)
98% 947
15.4 IPS
 
Pose Estimation (SP)
100% 1230
1.43 IPS
 
Pose Estimation (HP)
100% 1293
1.51 IPS
 
Pose Estimation (Q)
84% 1936
2.30 IPS
 
Object Detection (SP)
100% 278
22.0 IPS
 
Object Detection (HP)
100% 218
17.3 IPS
 
Object Detection (Q)
83% 820
66.3 IPS
 
Face Detection (SP)
100% 864
10.3 IPS
 
Face Detection (HP)
100% 864
10.3 IPS
 
Face Detection (Q)
95% 1938
23.1 IPS
 
Depth Estimation (SP)
100% 1183
9.11 IPS
 
Depth Estimation (HP)
99% 1174
9.04 IPS
 
Depth Estimation (Q)
64% 1879
17.2 IPS
 
Style Transfer (SP)
89% 2172
2.82 IPS
 
Style Transfer (HP)
89% 2253
2.92 IPS
 
Style Transfer (Q)
98% 4688
6.04 IPS
 
Image Super-Resolution (SP)
100% 533
19.7 IPS
 
Image Super-Resolution (HP)
100% 638
23.6 IPS
 
Image Super-Resolution (Q)
97% 1061
39.3 IPS
 
Text Classification (SP)
100% 344
459.8 IPS
 
Text Classification (HP)
100% 339
452.5 IPS
 
Text Classification (Q)
88% 728
981.8 IPS
 
Machine Translation (SP)
100% 467
8.04 IPS
 
Machine Translation (HP)
100% 554
9.55 IPS
 
Machine Translation (Q)
50% 432
15.4 IPS
 

Compare

Set Baseline Geekbench 6 now available