Geekbench AI Score

6074
Single Precision Score
37509
Half Precision Score
50921
Quantized Score
Geekbench AI 1.4.0 for macOS AArch64

Result Information

User Dachaz
Upload Date September 17 2025 08:07 PM
Views 5

System Information

AI Information
Framework Core ML
Backend Neural Engine
Device Apple M4 Max
System Information
Operating System macOS 26.0 (Build 25A354)
Model Mac16,9
Model IDMac16,9
Motherboard Mac16,9
CPU Information
Name Apple M4 Max
Topology 1 Processor, 16 Cores
Identifier Apple M4 Max
Base Frequency 4.51 GHz
Cluster 112 Cores
Cluster 24 Cores
Memory Information
Size 128.00 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 2199
408.9 IPS
 
Image Classification (HP)
100% 17499
3.25 KIPS
 
Image Classification (Q)
97% 21779
4.06 KIPS
 
Image Segmentation (SP)
100% 2530
41.0 IPS
 
Image Segmentation (HP)
100% 22601
366.4 IPS
 
Image Segmentation (Q)
99% 32318
523.9 IPS
 
Pose Estimation (SP)
100% 18611
21.7 IPS
 
Pose Estimation (HP)
100% 124157
144.9 IPS
 
Pose Estimation (Q)
95% 235003
275.4 IPS
 
Object Detection (SP)
100% 2734
216.9 IPS
 
Object Detection (HP)
100% 19470
1.54 KIPS
 
Object Detection (Q)
91% 24799
1.98 KIPS
 
Face Detection (SP)
100% 5379
63.9 IPS
 
Face Detection (HP)
100% 44978
534.4 IPS
 
Face Detection (Q)
100% 64518
766.6 IPS
 
Depth Estimation (SP)
100% 9097
70.1 IPS
 
Depth Estimation (HP)
99% 111859
861.8 IPS
 
Depth Estimation (Q)
85% 183678
1.44 KIPS
 
Style Transfer (SP)
100% 31658
40.7 IPS
 
Style Transfer (HP)
100% 234757
301.8 IPS
 
Style Transfer (Q)
96% 324481
418.7 IPS
 
Image Super-Resolution (SP)
100% 7142
263.7 IPS
 
Image Super-Resolution (HP)
100% 49631
1.83 KIPS
 
Image Super-Resolution (Q)
99% 77013
2.84 KIPS
 
Text Classification (SP)
100% 3254
4.34 KIPS
 
Text Classification (HP)
99% 5872
7.84 KIPS
 
Text Classification (Q)
93% 5711
7.67 KIPS
 
Machine Translation (SP)
100% 6712
115.6 IPS
 
Machine Translation (HP)
100% 16754
288.6 IPS
 
Machine Translation (Q)
98% 16896
292.0 IPS
 

Compare

Set Baseline Geekbench 6 now available