Google Cloud GPU Pricing
All Google Cloud GPU instances and pricing. 16 GPU models from $0.73/hr.
Hyperscaler with NVIDIA GPUs and TPU pod slices across Google regions; no public stock API. We track 16 GPU models across 820 configurations. Google Cloud (GCP) lists hourly on-demand rates; the provider publishes no stock feed, so availability shows as unknown. Egress is $0.12/GB — budget for it on data-heavy jobs.
Egress Pricing
30-day reliability depth
GPU Finder does not treat availability as a static yes/no flag. For Google Cloud, current stock badges are paired with a 30-day reliability score based on our hourly stock checks — the share of tracked time each listing was reported available. How we compute this →
- 0
- listings with a 30-day score
- 0%
- of listings on this page have a score
- —
- longest stock history on this page
This page has live pricing and current stock rows, but not enough retained availability history yet to publish 30-day reliability percentages.
Caveat: scores stay hidden until a listing has at least 48 tracked hours, and some provider APIs expose coarse capacity levels instead of exact stock counts. Use the score as historical depth beside current availability, not as a guarantee that a GPU will still be allocatable when you click through.
Google Cloud GPU Pricing Guide
Use this Google Cloud GPU pricing page to check what GCP charges per hour for Compute Engine GPU VMs — A3 (H100/H200), A4 (B200), A2 (A100), G2 (L4) and N1 (T4) — across the live rows GPU Finder tracks, with Spot VM prices where Google publishes them. The table below is the source of truth for hourly rates; these shortcuts jump from Google Cloud's price into the same GPU across every provider.
How much does Google Cloud GPU pricing start at?
In GPU Finder's current Google Cloud data, single-GPU on-demand pricing starts at $0.73/hr and runs up to $26.90/hr; the full table lists 820 configurations, including larger multi-GPU nodes priced above that range. Compute Engine bills GPU VMs per second after a one-minute minimum.
What does a GCP H100 cost per hour?
The cheapest Google Cloud H100 listing right now is 8x H100 at $82.93/hr on demand, i.e. $10.37 per GPU-hour. As a Spot VM the same configuration drops to $6.21 per GPU-hour, but Google can reclaim it at any time.
What do the Google Cloud A100 and L4 cost per hour?
The cheapest Google Cloud A100 listing right now is 8x A100 at $28.03/hr on demand, i.e. $3.50 per GPU-hour. For inference, the L4 starts at $1.13 per GPU-hour.
Does Google Cloud charge for GPU egress?
Yes. Internet egress from Google Cloud is about $0.12/GB at the rate GPU Finder tracks, so moving a 100 GB checkpoint out costs roughly $12.00 on top of GPU hours. Free-egress clouds such as Lambda avoid that line item entirely.
Can I compare Google Cloud TPU and GPU prices here?
Yes. GPU Finder lists Google Cloud TPUs next to its NVIDIA GPUs; the cheapest TPU listing right now is TPU v5 Lite (tpu-v5litepod-1 slice) at $1.96/hr for the whole slice. A slice can hold one chip or many, so compare them against GPUs by throughput on your model rather than per device, and only if your code runs on JAX/XLA or PyTorch/XLA.
Is Google Cloud cheaper than AWS for GPUs?
On list on-demand prices the two hyperscalers are close and the gap swings by GPU and region; specialist clouds such as Lambda, Runpod and Nebius usually undercut both. Open the AWS vs Google Cloud or Google Cloud vs Nebius comparison for the live per-GPU numbers on every model both list.