Find the right GCP Compute Engine machine type by vCPU, memory, and workload.
Last verified: May 2026
Output will appear here...The GCP Compute Instance Finder helps you navigate Google Cloud's machine type catalog to find the right Compute Engine instance for your workload. Filter by vCPU count, memory, GPU availability, and machine family to find matching options with pricing estimates. The tool covers all current machine families including general purpose (E2, N2, N2D, C3), compute optimized (C2, C2D, H3), memory optimized (M1, M2, M3), and accelerator-optimized (A2, G2).
GCP machine types follow the pattern [family]-[type]-[vCPUs]. For example, e2-standard-8 means E2 family, standard type (balanced CPU-to-memory), 8 vCPUs. Types include: standard (4 GB/vCPU), highcpu (1 GB/vCPU), highmem (8 GB/vCPU). Custom types allow arbitrary CPU and memory combinations.
E2 instances are the most cost-effective for general workloads and automatically use a mix of Intel and AMD processors. N2 instances offer consistent Intel Cascade Lake performance, higher sustained-use discounts, and support for local SSDs and sole-tenant nodes. Choose E2 for cost savings and N2 for predictable performance.
Sustained-use discounts (SUDs) automatically reduce pricing by up to 30% as instances run longer in a month. Committed-use discounts (CUDs) offer 1-year (37% off) or 3-year (55% off) savings commitments on vCPU and memory. CUDs apply at the project level and work across machine types. E2 instances receive SUDs but not CUDs.
Your team needs 20 VMs for a Kubernetes node pool sized for general-purpose web traffic (4 vCPU, 16 GB RAM). The finder compares: e2-standard-4 at $97/month (no SUDs but cheapest list), n2-standard-4 at $108/month list dropping to ~$76/month with SUDs, n2d-standard-4 (AMD) at ~$66/month with SUDs. You pick n2d-standard-4 with a 1-year CUD = $51/month effective. Compared to the e2-standard-4 default, savings: $46/month × 20 nodes = $11K/year.
The finder filters GCE machine types by vCPU/memory/family/GPU criteria and returns matches sorted by effective monthly cost (after auto-applied SUDs for eligible families). Each result includes the on-demand rate, the SUDs-adjusted rate (for full-month operation), and the 1-year and 3-year CUD rates so you can model commitment scenarios.
C3 and C3D (the latest compute-optimized families) deliver 20-30% better price-performance than older C2/C2D for compute-bound workloads. C3 (Intel Sapphire Rapids) and C3D (AMD Genoa) are typically a drop-in replacement and well worth the migration effort for batch processing, video encoding, and HPC.
Custom machine types (e.g., custom-6-49152 for 6 vCPU + 48 GB RAM) avoid the rigid CPU-to-memory ratios of standard types. A workload needing 4 vCPU and 24 GB RAM would otherwise force you into n2-standard-8 (8 vCPU, 32 GB RAM, $$$) — a custom type can be 30% cheaper for that exact spec.
T2D (AMD Tau) and T2A (Ampere ARM) families don't support sustained-use discounts but have the lowest list prices in their tiers. For workloads running 24/7, the SUDs on N2/N2D often make those families cheaper despite higher list price — always compare effective rate, not list rate.
Was this tool helpful?
Disclaimer: This tool runs entirely in your browser. No data is sent to our servers. Always verify outputs before using them in production. AWS, Azure, and GCP are trademarks of their respective owners.