Real-World Performance

Workload Benchmarks

How Google Cloud VMs perform on actual workloads -- not just synthetic tests. Find the best VM for your specific use case.

10 benchmarks

Web Serving

Static file serving, API throughput, and request handling performance for web applications.

Top pick

C3 series

Maximum throughput and lowest latency

10 benchmarks

Database

Transactional throughput, query performance, and in-memory operations for database workloads.

Top pick

Z3 series

I/O-intensive OLTP databases

7 benchmarks

ML Inference

Model inference speed for LLMs, image classification, and generative AI workloads.

Top pick

G2 series

Cost-effective GPU inference for most models

4 benchmarks

ML Training

Training throughput for deep learning models using GPU-accelerated compute.

Top pick

A3 series

Large-scale model training and fine-tuning

6 benchmarks

Video Encoding

Video transcoding and encoding performance for media processing pipelines.

Top pick

G2 series

High-throughput hardware encoding for streaming

7 benchmarks

HPC

High-performance computing benchmarks including LINPACK and CFD simulations.

Top pick

H3 series

Tightly-coupled HPC and MPI workloads

Best VM by Workload -- Quick Reference

WorkloadBest PerformanceBest ValueBudget Pick
Web ServingC3 (c3-standard-8)N2 (n2-standard-8)E2 (e2-standard-4)
DatabaseZ3 (z3-standard-88)M3 (m3-ultramem-32)N2 (n2-standard-8)
ML InferenceG2 (g2-standard-8)A3 (a3-highgpu-8g)G2 (g2-standard-4)
ML TrainingA3 (a3-highgpu-8g)G2 (g2-standard-96)G2 (g2-standard-24)
Video EncodingG2 (g2-standard-8)H3 (h3-standard-88)C3D (c3d-standard-8)
HPCH3 (h3-standard-88)C3D (c3d-standard-60)N2 (n2-standard-32)