Google Tensor G5

Google Tensor G5
Benchmarks & Specs

Last updated:
The Google Tensor G5 marks a significant advancement in the world of mobile processors. This chip, manufactured using the advanced 3-nanometer process, demonstrates remarkable energy efficiency. This innovative manufacturing significantly contributes to extended battery life. It also enables a high density of transistors for more performance.

One of the greatest strengths of the Google Tensor G5 lies in its outstanding artificial intelligence capabilities. It was specifically developed for complex AI operations and on-device machine learning. This allows sophisticated algorithms to be processed directly and efficiently. This leads to faster and more secure applications without cloud connectivity.

The integrated graphics unit, known as IMG DXT-48-1536, provides a compelling visual representation. It offers solid graphics performance for everyday use. It also handles more demanding visual tasks such as editing photos or playing high-resolution videos with ease. Users experience consistently smooth display in apps and media content.

Thanks to a high memory bandwidth of 51 GB/s, data is processed very quickly. This ensures a consistently responsive system in all situations. The overall system performance benefits noticeably from this and ensures a smooth user experience. The Google Tensor G5 is a robust choice for a variety of applications in modern smartphones and tablets.

It reliably delivers high and efficient performance. We found the integration of the components to be particularly successful. The design of this system is clearly aimed at a seamless and intuitive user experience. The underlying ARM architecture with the code name G5 Laguna forms the technological basis for this.
  • Manufactured using the 3-nanometer process
  • Outstanding AI performance
  • Integrated graphics unit IMG DXT-48-1536
  • High memory bandwidth of 51 GB/s

Performance Overview
Average performance across multiple benchmarks

Single-core performance
Rank 13 / 176
In segment Smartphone / Tablet
Multi-core performance
Rank 13 / 177
In segment Smartphone / Tablet
Geekbench 6 Single-Core

Geekbench 6 Single-Core
Single-core performance

Apple A15 Bionic (4-GPU)
6C / 6T · 3.23 GHz
2,370
Apple A15 Bionic (5-GPU)
6C / 6T · 3.23 GHz
2,370
Google Tensor G5
8C / 8T · 3.78 GHz
2,362
Qualcomm Snapdragon 8 Gen 3
8C / 8T · 3.40 GHz
2,293
MediaTek Dimensity 9300
8C / 8T · 3.25 GHz
2,234
Geekbench 6 Multi-Core

Geekbench 6 Multi-Core
Multi-core performance

Qualcomm Snapdragon 8 Gen 3
8C / 8T · 3.40 GHz
7,117
Samsung Exynos 2400
10C / 10T · 3.21 GHz
6,971
Google Tensor G5
8C / 8T · 3.78 GHz
6,403
Apple A16 Bionic
6C / 6T · 3.46 GHz
6,299
Qualcomm Snapdragon Microsoft SQ3
8C / 8T · 3.00 GHz
6,133
iGPU - FP32 Performance (Single-precision GFLOPS)

iGPU - FP32 Performance (Single-precision GFLOPS)

Qualcomm Snapdragon 888+
Qualcomm Adreno 660 AV1
1,720
Apple A15 Bionic (5-GPU)
Apple A15 (5 GPU Cores)
1,713
Google Tensor G5
IMG DXT-48-1536
1,690
Qualcomm Snapdragon 8c
Qualcomm Adreno 675
1,550
Apple A16 Bionic (5-CPU 4-GPU)
Apple A16 (4 GPU Cores)
1,431

More benchmarks

At a glance

PropertyValue
FamilyGoogle Tensor (5)
CPU groupGoogle Tensor G5 (1)
ArchitectureG5 Laguna
Technology3 nm
SegmentSmartphone / Tablet
Socket
Generation5
PredecessorGoogle Tensor G4
Q4/2024
Successor

CPU Cores and Base Frequency

PropertyValue
CPU Cores / Threads8 / 8
Hyperthreading / SMT
Core architecturehybrid (Prime / big.LITTLE)
Core Cluster 1: 1x Cortex-X4
3.78 GHz
Core Cluster 2: 5x Cortex-A725
3.05 GHz
Core Cluster 3: 2x Cortex-A520
2.25 GHz
L2-Cache
L3-Cache
OverclockingNo

Integrated graphics (iGPU)

PropertyValue
GPU nameIMG DXT-48-1536
GPU frequency0.40 - 1.10 GHz
CUs / Shader6 / 1536
Raytracing
Max. displays0
Max. GPU Memory8 GB
Technology3 nm
Release dateQ4/2025

NPU AI performance

PropertyValue
AI hardwareGoogle Tensor AI
AI specificationsGoogle Edge TPU
NPU + CPU + iGPU

Memory & PCIe

Memory typeMemory bandwidth
LPDDR5-6400
51.2 GB/s
PropertyValue
Max. Memory16 GB
Memory channels4
ECC
PCIe
PCIe Bandwidth

Thermal Management

PropertyValue
TDP10 W
TDP (PL2)
TDP up
TDP down
T. junction max.

Technical details

PropertyValue
Chip designChiplet
AES-NI
Operating systemsAndroid
Instruction setArmv8.7-A (64 bit)
ISA extensions
Release dateQ3/2025
Release price
Documents
Google Tensor G5
Google Tensor G5
Buy now on Amazon and save! Buy on Amazon