Geekbench AI Score

3431
Single Precision Score
21182
Half Precision Score
22866
Quantized Score
Geekbench AI 1.0.0 Pro for macOS AArch64

Result Information

User Swissnightlover
Upload Date August 19 2024 05:05 PM
Views 16

System Information

AI Information
Framework Core ML
Backend Neural Engine
Device Apple M2
System Information
Operating System macOS 15.1 (Build 24B5024e)
Model MacBook Air (2022)
Model IDMac14,2
Motherboard Mac14,2
CPU Information
Name Apple M2
Topology 1 Processor, 8 Cores
Identifier Apple M2
Base Frequency 3.47 GHz
Cluster 14 Cores
Cluster 24 Cores
Memory Information
Size 8.00 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 1366
254.0 IPS
 
Image Classification (HP)
100% 14170
2.64 KIPS
 
Image Classification (Q)
99% 14294
2.67 KIPS
 
Image Segmentation (SP)
100% 1516
24.6 IPS
 
Image Segmentation (HP)
100% 9104
147.6 IPS
 
Image Segmentation (Q)
100% 9004
146.0 IPS
 
Pose Estimation (SP)
100% 6837
7.98 IPS
 
Pose Estimation (HP)
100% 100504
117.3 IPS
 
Pose Estimation (Q)
94% 100323
117.7 IPS
 
Object Detection (SP)
100% 1544
122.5 IPS
 
Object Detection (HP)
99% 14928
1.19 KIPS
 
Object Detection (Q)
91% 16993
1.36 KIPS
 
Face Detection (SP)
100% 3095
36.8 IPS
 
Face Detection (HP)
100% 23337
277.3 IPS
 
Face Detection (Q)
100% 35965
427.3 IPS
 
Depth Estimation (SP)
100% 5643
43.5 IPS
 
Depth Estimation (HP)
99% 74854
576.7 IPS
 
Depth Estimation (Q)
97% 81751
632.1 IPS
 
Style Transfer (SP)
100% 18683
24.0 IPS
 
Style Transfer (HP)
100% 154980
199.2 IPS
 
Style Transfer (Q)
100% 156430
201.1 IPS
 
Image Super-Resolution (SP)
100% 4068
150.2 IPS
 
Image Super-Resolution (HP)
100% 40116
1.48 KIPS
 
Image Super-Resolution (Q)
98% 46149
1.71 KIPS
 
Text Classification (SP)
100% 2190
2.92 KIPS
 
Text Classification (HP)
100% 3662
4.89 KIPS
 
Text Classification (Q)
93% 3582
4.81 KIPS
 
Machine Translation (SP)
100% 3560
61.3 IPS
 
Machine Translation (HP)
100% 2363
40.7 IPS
 
Machine Translation (Q)
98% 2343
40.5 IPS
 

Compare

Set Baseline Geekbench 6 now available