Geekbench AI Score

29395
Single Precision Score
46258
Half Precision Score
43197
Quantized Score
Geekbench AI 1.7.0 for macOS AArch64

Result Information

User gundar
Upload Date March 29 2026 04:10 AM
Views 23

System Information

AI Information
Framework Core ML
Backend GPU
Device Apple M5 Max
System Information
Operating System macOS 26.4 (Build 25E246)
Model Mac17,6
Model IDMac17,6
Motherboard Mac17,6
CPU Information
Name Apple M5 Max
Topology 1 Processor, 18 Cores
Identifier Apple M5 Max
Base Frequency 4.61 GHz
Cluster 16 Cores
Cluster 212 Cores
Memory Information
Size 128.00 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 10470
1.95 KIPS
 
Image Classification (HP)
100% 13449
2.50 KIPS
 
Image Classification (Q)
97% 12068
2.25 KIPS
 
Image Segmentation (SP)
100% 27959
453.2 IPS
 
Image Segmentation (HP)
100% 42232
684.6 IPS
 
Image Segmentation (Q)
99% 38838
629.6 IPS
 
Pose Estimation (SP)
100% 130462
152.2 IPS
 
Pose Estimation (HP)
100% 335253
391.2 IPS
 
Pose Estimation (Q)
95% 289371
339.2 IPS
 
Object Detection (SP)
100% 10047
796.9 IPS
 
Object Detection (HP)
100% 13697
1.09 KIPS
 
Object Detection (Q)
94% 14018
1.12 KIPS
 
Face Detection (SP)
100% 49342
586.3 IPS
 
Face Detection (HP)
100% 72188
857.7 IPS
 
Face Detection (Q)
100% 64749
769.4 IPS
 
Depth Estimation (SP)
100% 53140
409.4 IPS
 
Depth Estimation (HP)
99% 86576
667.0 IPS
 
Depth Estimation (Q)
86% 82823
646.6 IPS
 
Style Transfer (SP)
100% 264533
340.1 IPS
 
Style Transfer (HP)
100% 547623
704.0 IPS
 
Style Transfer (Q)
96% 598678
772.6 IPS
 
Image Super-Resolution (SP)
100% 34776
1.28 KIPS
 
Image Super-Resolution (HP)
100% 52130
1.92 KIPS
 
Image Super-Resolution (Q)
99% 42445
1.57 KIPS
 
Text Classification (SP)
99% 6335
8.46 KIPS
 
Text Classification (HP)
100% 10848
14.5 KIPS
 
Text Classification (Q)
93% 10202
13.7 KIPS
 
Machine Translation (SP)
100% 8217
141.5 IPS
 
Machine Translation (HP)
100% 8888
153.1 IPS
 
Machine Translation (Q)
98% 8559
147.9 IPS
 

Compare

Set Baseline Geekbench AI now available