# Google Compute Engine (Linux VMs)

Official pricing: https://cloud.google.com/products/compute/pricing/general-purpose  
Category: hyperscaler · Isolation: vm

## Pricing regimes (raw)

- **On-demand E2 / N2D / C3D (us-central1)** (sizes): c3d-standard-4 4 vCPU/16 GiB $0.181596/h; c3d-standard-8 8 vCPU/32 GiB $0.363192/h; c3d-standard-16 16 vCPU/64 GiB $0.726384/h; c3d-standard-30 30 vCPU/120 GiB $1.36197/h; c3d-standard-60 60 vCPU/240 GiB $2.72394/h; c3d-standard-90 90 vCPU/360 GiB $4.08591/h
- **On-demand Tau T2A Arm (us-central1)** (sizes): t2a-standard-1 1 vCPU/4 GiB $0.0385/h; t2a-standard-2 2 vCPU/8 GiB $0.077/h; t2a-standard-4 4 vCPU/16 GiB $0.154/h; t2a-standard-8 8 vCPU/32 GiB $0.308/h; t2a-standard-16 16 vCPU/64 GiB $0.616/h; t2a-standard-32 32 vCPU/128 GiB $1.232/h
- **Spot E2 / N2D / C3D (us-central1)** (sizes): c3d-standard-4 4 vCPU/16 GiB $0.048728/h; c3d-standard-8 8 vCPU/32 GiB $0.097456/h; c3d-standard-16 16 vCPU/64 GiB $0.194912/h; c3d-standard-30 30 vCPU/120 GiB $0.36546/h; c3d-standard-60 60 vCPU/240 GiB $0.73092/h; c3d-standard-90 90 vCPU/360 GiB $1.09638/h [spot]
- **Spot Tau T2A Arm (us-central1)** (sizes): t2a-standard-1 1 vCPU/4 GiB $0.0231/h; t2a-standard-2 2 vCPU/8 GiB $0.0462/h; t2a-standard-4 4 vCPU/16 GiB $0.0924/h; t2a-standard-8 8 vCPU/32 GiB $0.1848/h; t2a-standard-16 16 vCPU/64 GiB $0.3696/h; t2a-standard-32 32 vCPU/128 GiB $0.7392/h [spot]
- **Compute Flexible CUD 1-year (E2 / N2D / C3D)** (sizes): c3d-standard-4 4 vCPU/16 GiB $0.13074912/h; c3d-standard-8 8 vCPU/32 GiB $0.26149824/h; c3d-standard-16 16 vCPU/64 GiB $0.52299648/h; c3d-standard-30 30 vCPU/120 GiB $0.9806184/h; c3d-standard-60 60 vCPU/240 GiB $1.9612368/h; c3d-standard-90 90 vCPU/360 GiB $2.9418552/h [commit]
- **Compute Flexible CUD 3-year (E2 / N2D / C3D)** (sizes): c3d-standard-4 4 vCPU/16 GiB $0.09806184/h; c3d-standard-8 8 vCPU/32 GiB $0.19612368/h; c3d-standard-16 16 vCPU/64 GiB $0.39224736/h; c3d-standard-30 30 vCPU/120 GiB $0.7354638/h; c3d-standard-60 60 vCPU/240 GiB $1.4709276/h; c3d-standard-90 90 vCPU/360 GiB $2.2063914/h [commit]

## Features

Yes: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, full VM (own kernel), Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, GPU, arm64, HTTP method/path egress rules, live fork (no pause), audit logs, EU data residency, snapshots on demand, gVisor or VM (no shared kernel), inbound access rules, firewall inside (nftables), extra volumes, shared volumes

No: memory snapshots, idle auto-stop, nested virtualization, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, preinstalled agents, HTTPS preview URLs, self-hosting, BYOC, open source, Windows, macOS, Python, Node.js, wake on request, live resize, memory fork, webhooks, spend limits, MCP server, automatic snapshots, agent harness API, hosted agent API (their own agent), secret proxy, secret proxy for any API

Unknown: everything else. Evidence (source + quote) per feature: https://battleships.dev/data/providers/gcp-compute.json → feature_evidence

## Caveats

- DIY VM: you manage images, SSH, scaling and lifecycle; no sandbox API. Boot time not measured here (boot_overhead_s unset).
- All prices are us-central1 (Iowa), the default region of the official tables; other regions cost more. Predefined shapes only (custom machine types, N2/N4/C3/C4 and shared-core e2-micro/small/medium not encoded).
- Billed per second after a 1-minute minimum while RUNNING; a stopped VM bills only its disks. Boot disk not included in the machine price (balanced PD ~$0.10/GiB-month, carried in storage).
- Sustained-use discounts (N2D/C3D only, up to 20% for full-month use) not modelled; E2 has none.
- Disk, snapshot, egress and IPv4 rates copied from the gcp-windows card (same provider, verified 2026-09-28): balanced PD $0.10/GiB-month, standard snapshot $0.05/GiB-month, Premium Tier egress $0.12/GiB after 1 GiB, external IPv4 $0.005/h.
- nested_virt=false: every priced mode is E2 / N2D / C3D (AMD) or T2A (Arm); GCP docs exclude "E2 VMs", "VMs powered by ARM processors" and "VMs powered by AMD processors except N4D" from nested virtualization. Intel N2/C3/N4 etc. support it but are not priced here.
- Only a subset of machine families is encoded. Not encoded: N4/N4A/N4D/N2/N1/C4/C4A/C4D/C3, memory/compute/accelerator-optimized families, custom machine types, shared-core E2, sole-tenant, confidential VMs, reservations, Windows/RHEL/SLES image surcharges and GPUs (e.g. c4-standard-2, 2 vCPU/7 GiB, $0.096866/h in Iowa).