Geekbench AI Score

940
Single Precision Score
989
Half Precision Score
653
Quantized Score
Geekbench AI 1.4.0 for Windows AVX2

Result Information

User cas1
Upload Date September 02 2025 01:53 PM
Views 9

System Information

AI Information
Framework ONNX
Backend DirectML
Device Intel(R) UHD Graphics
System Information
Operating System Microsoft Windows 11 Home (64-bit)
Model LENOVO 83LY
Motherboard LENOVO LNVNB161216
Power PlanBalanced
CPU Information
Name Intel Core i7-13650HX
Topology 1 Processor, 14 Cores, 20 Threads
Identifier GenuineIntel Family 6 Model 183 Stepping 1
Base Frequency 2.60 GHz
Cluster 16 Cores
Cluster 28 Cores
Memory Information
Size 32.00 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 402
74.7 IPS
 
Image Classification (HP)
100% 397
73.9 IPS
 
Image Classification (Q)
100% 335
62.3 IPS
 
Image Segmentation (SP)
100% 535
8.67 IPS
 
Image Segmentation (HP)
100% 602
9.76 IPS
 
Image Segmentation (Q)
98% 418
6.79 IPS
 
Pose Estimation (SP)
100% 2668
3.11 IPS
 
Pose Estimation (HP)
100% 2925
3.41 IPS
 
Pose Estimation (Q)
96% 1919
2.25 IPS
 
Object Detection (SP)
100% 444
35.2 IPS
 
Object Detection (HP)
98% 378
30.1 IPS
 
Object Detection (Q)
86% 373
30.0 IPS
 
Face Detection (SP)
100% 880
10.5 IPS
 
Face Detection (HP)
100% 939
11.2 IPS
 
Face Detection (Q)
97% 673
8.02 IPS
 
Depth Estimation (SP)
100% 1622
12.5 IPS
 
Depth Estimation (HP)
98% 1833
14.2 IPS
 
Depth Estimation (Q)
77% 1230
9.86 IPS
 
Style Transfer (SP)
100% 5416
6.96 IPS
 
Style Transfer (HP)
100% 7371
9.48 IPS
 
Style Transfer (Q)
98% 4159
5.36 IPS
 
Image Super-Resolution (SP)
100% 970
35.8 IPS
 
Image Super-Resolution (HP)
100% 1330
49.1 IPS
 
Image Super-Resolution (Q)
99% 530
19.6 IPS
 
Text Classification (SP)
100% 498
664.5 IPS
 
Text Classification (HP)
100% 363
485.0 IPS
 
Text Classification (Q)
97% 277
371.5 IPS
 
Machine Translation (SP)
100% 572
9.85 IPS
 
Machine Translation (HP)
100% 558
9.60 IPS
 
Machine Translation (Q)
70% 280
5.25 IPS
 

Compare

Set Baseline Geekbench 6 now available