Geekbench AI Score

4835
Single Precision Score
28415
Half Precision Score
32334
Quantized Score
Geekbench AI 1.7.0 for macOS AArch64

Result Information

Upload Date November 21 2026 05:12 AM
Views 4

System Information

AI Information
Framework Core ML
Backend Neural Engine
Device Apple M2 Ultra
System Information
Operating System macOS 27.0.1 (Build 26A434)
Model Mac Studio (2023)
Model IDMac14,14
Motherboard Mac14,14
CPU Information
Name Apple M2 Ultra
Topology 1 Processor, 24 Cores
Identifier Apple M2 Ultra
Base Frequency 3.48 GHz
Cluster 116 Cores
Cluster 28 Cores
Memory Information
Size 64.00 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 1407
261.7 IPS
 
Image Classification (HP)
100% 13711
2.55 KIPS
 
Image Classification (Q)
97% 17875
3.34 KIPS
 
Image Segmentation (SP)
100% 1740
28.2 IPS
 
Image Segmentation (HP)
100% 19492
316.0 IPS
 
Image Segmentation (Q)
99% 25104
407.0 IPS
 
Pose Estimation (SP)
100% 18204
21.2 IPS
 
Pose Estimation (HP)
100% 104662
122.1 IPS
 
Pose Estimation (Q)
95% 103870
121.7 IPS
 
Object Detection (SP)
100% 1791
142.0 IPS
 
Object Detection (HP)
99% 15381
1.22 KIPS
 
Object Detection (Q)
89% 16789
1.34 KIPS
 
Face Detection (SP)
100% 3615
43.0 IPS
 
Face Detection (HP)
100% 36821
437.5 IPS
 
Face Detection (Q)
100% 49437
587.4 IPS
 
Depth Estimation (SP)
100% 6127
47.2 IPS
 
Depth Estimation (HP)
99% 84708
652.6 IPS
 
Depth Estimation (Q)
85% 110413
863.7 IPS
 
Style Transfer (SP)
100% 32382
41.6 IPS
 
Style Transfer (HP)
100% 218695
281.1 IPS
 
Style Transfer (Q)
96% 201830
260.5 IPS
 
Image Super-Resolution (SP)
100% 5450
201.3 IPS
 
Image Super-Resolution (HP)
100% 41028
1.51 KIPS
 
Image Super-Resolution (Q)
99% 47791
1.76 KIPS
 
Text Classification (SP)
100% 4839
6.46 KIPS
 
Text Classification (HP)
99% 7398
9.87 KIPS
 
Text Classification (Q)
93% 7427
9.97 KIPS
 
Machine Translation (SP)
100% 4629
79.7 IPS
 
Machine Translation (HP)
100% 3853
66.4 IPS
 
Machine Translation (Q)
98% 4083
70.6 IPS
 

Compare

Set Baseline Geekbench AI now available