Geekbench AI Score

4154
Single Precision Score
30814
Half Precision Score
34583
Quantized Score
Geekbench AI 1.3.0 for macOS AArch64

Result Information

Upload Date April 06 2025 08:26 AM
Views 2

System Information

AI Information
Framework Core ML
Backend Neural Engine
Device Apple M3 Pro
System Information
Operating System macOS 15.4 (Build 24E248)
Model MacBook Pro (14-inch, Nov 2023)
Model IDMac15,6
Motherboard Mac15,6
CPU Information
Name Apple M3 Pro
Topology 1 Processor, 11 Cores
Identifier Apple M3 Pro
Base Frequency 4.05 GHz
Cluster 15 Cores
Cluster 26 Cores
Memory Information
Size 18.00 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 1624
302.0 IPS
 
Image Classification (HP)
100% 16505
3.07 KIPS
 
Image Classification (Q)
97% 19011
3.55 KIPS
 
Image Segmentation (SP)
100% 1830
29.7 IPS
 
Image Segmentation (HP)
100% 22535
365.3 IPS
 
Image Segmentation (Q)
99% 27968
453.4 IPS
 
Pose Estimation (SP)
100% 9018
10.5 IPS
 
Pose Estimation (HP)
100% 120278
140.3 IPS
 
Pose Estimation (Q)
95% 119915
140.5 IPS
 
Object Detection (SP)
100% 1819
144.3 IPS
 
Object Detection (HP)
100% 17255
1.37 KIPS
 
Object Detection (Q)
91% 20085
1.60 KIPS
 
Face Detection (SP)
100% 3794
45.1 IPS
 
Face Detection (HP)
100% 39932
474.5 IPS
 
Face Detection (Q)
100% 50221
596.7 IPS
 
Depth Estimation (SP)
100% 7148
55.1 IPS
 
Depth Estimation (HP)
99% 97070
747.9 IPS
 
Depth Estimation (Q)
85% 121155
947.2 IPS
 
Style Transfer (SP)
100% 22295
28.7 IPS
 
Style Transfer (HP)
100% 226324
290.9 IPS
 
Style Transfer (Q)
96% 219340
283.0 IPS
 
Image Super-Resolution (SP)
100% 5246
193.7 IPS
 
Image Super-Resolution (HP)
100% 44797
1.65 KIPS
 
Image Super-Resolution (Q)
99% 54506
2.01 KIPS
 
Text Classification (SP)
100% 2850
3.80 KIPS
 
Text Classification (HP)
100% 5122
6.84 KIPS
 
Text Classification (Q)
93% 5175
6.94 KIPS
 
Machine Translation (SP)
100% 3472
59.8 IPS
 
Machine Translation (HP)
100% 4967
85.6 IPS
 
Machine Translation (Q)
98% 5077
87.7 IPS
 

Compare

Set Baseline Geekbench 6 now available