Geekbench AI Score

2608
Single Precision Score
16079
Half Precision Score
18073
Quantized Score
Geekbench AI 1.1.0 for macOS AArch64

Result Information

Upload Date November 18 2024 03:00 PM
Views 3

System Information

AI Information
Framework Core ML
Backend Neural Engine
Device Apple M3 Max
System Information
Operating System macOS 14.6.1 (Build 23G93)
Model MacBook Pro (16-inch, Nov 2023)
Model IDMac15,9
Motherboard Mac15,9
CPU Information
Name Apple M3 Max
Topology 1 Processor, 16 Cores
Identifier Apple M3 Max
Base Frequency 4.05 GHz
Cluster 112 Cores
Cluster 24 Cores
Memory Information
Size 128.00 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 1143
212.5 IPS
 
Image Classification (HP)
100% 7148
1.33 KIPS
 
Image Classification (Q)
97% 7663
1.43 KIPS
 
Image Segmentation (SP)
100% 812
13.2 IPS
 
Image Segmentation (HP)
100% 6459
104.7 IPS
 
Image Segmentation (Q)
99% 12751
206.7 IPS
 
Pose Estimation (SP)
100% 14545
17.0 IPS
 
Pose Estimation (HP)
100% 103237
120.5 IPS
 
Pose Estimation (Q)
95% 100803
118.1 IPS
 
Object Detection (SP)
100% 1196
94.9 IPS
 
Object Detection (HP)
100% 10037
796.1 IPS
 
Object Detection (Q)
93% 10020
799.5 IPS
 
Face Detection (SP)
100% 2198
26.1 IPS
 
Face Detection (HP)
100% 12644
150.2 IPS
 
Face Detection (Q)
100% 14752
175.3 IPS
 
Depth Estimation (SP)
100% 4223
32.5 IPS
 
Depth Estimation (HP)
99% 61317
472.4 IPS
 
Depth Estimation (Q)
85% 70792
553.5 IPS
 
Style Transfer (SP)
100% 15736
20.2 IPS
 
Style Transfer (HP)
100% 149481
192.2 IPS
 
Style Transfer (Q)
96% 141994
183.2 IPS
 
Image Super-Resolution (SP)
100% 3384
125.0 IPS
 
Image Super-Resolution (HP)
100% 21236
784.1 IPS
 
Image Super-Resolution (Q)
99% 26589
981.8 IPS
 
Text Classification (SP)
100% 893
1.19 KIPS
 
Text Classification (HP)
100% 3625
4.84 KIPS
 
Text Classification (Q)
93% 3541
4.75 KIPS
 
Machine Translation (SP)
100% 2049
35.3 IPS
 
Machine Translation (HP)
100% 2707
46.6 IPS
 
Machine Translation (Q)
98% 2699
46.6 IPS
 

Compare

Set Baseline Geekbench 6 now available