Google Compute Engine GPU VMs
- Free allowance
- Not stated on https://cloud.google.com/products/compute/pricing/accelerator-optimized (read 2026-10-30): the page is a price sheet that bills every GPU machine type by the hour, for example a2-ultragpu-1g (one A100) at USD 5.06879789 an hour on demand, and prints no free allowance
- Verdict
- paidno free allowance; listed for its price
- Cost
- no price stated
- Identity
- cardcard required, no phone
- Binding limit
- write-gatedno free-tier allowance applies to any GPU or TPU; billing starts with the first second
- Evidence
- documented12 sources; a first-party page was read and the sentence quoted
- Freshness
- stale on 2026-12-04checked 2026-11-03 (live); stale on 2026-12-04
| Free tier | not stated on https://cloud.google.com/products/compute/pricing/accelerator-optimized (read 2026-10-30): the page is a price sheet that bills every GPU machine type by the hour, for example a2-ultragpu-1g (one A100) at USD 5.06879789 an hour on demand, and prints no free allowance unit none; metered on wall-clock; The Free Tier document says 'GPUs and TPUs are not included in the Free Tier offer.', and a non-billable Free Trial account cannot add GPUs to VM instances, so the USD 300 Welcome credit does not reach a GPU before the account is upgraded. After the upgrade to a Paid billing account the same document says GPUs are unlocked and the remaining credit stays usable on products and services covered by the Free Trial for the original 90 days from sign-up; its list of what the credit cannot pay for names Gemini API in AI Studio and managed-API partner models, not GPUs, and a Google Cloud Support reply states the credit is applied to GPU usage after the upgrade, so the credit pays for GPU time (inferred from that exclusion list and the support reply; no page prices a GPU against the credit). |
|---|---|
| Price | none stated |
| Limits |
|
| Identity | card Free Trial signup requires a credit card or other payment method, and the upgrade to a Paid billing account that GPU and Windows Server VMs need is made on that billing account |
| Egress | not stated |
| Per-object cap | not stated |
| Retention | not stated |
| Interface | unknown |
| Tooling | part of the text of Free tier, above |
| Composes with | part of the text of Free tier, above |
| Reliability | no reliability event recorded |
| Evidence note | documented both the accelerator-optimized rate card and the free-features document were read on 2026-10-22. The page proves the per-GPU rate and the free-features document proves there is no free GPU allowance |
| Parsed numbers |
|
Evidence 12
| date | source | what it proved |
|---|---|---|
| 2026-10-02 | cloud.google.com/products/compute/pricing/accelerator-optimi | not published |
| 2026-10-22 | cloud.google.com/products/compute/pricing/accelerator-optimi | 'GPUs and TPUs are not included in the Free Tier offer. You are always charged for GPUs and TPUs that you add to VM instances' (free-features document), against the accelerator-optimized rate card read the same day: a2-ultragpu-1g at $5.06879789 an hour on demand |
| 2026-10-30 | cloud.google.com/products/compute/pricing/accelerator-optimi | the accelerator-optimized price sheet bills a one-A100 a2-ultragpu-1g at USD 5.06879789 an hour on demand; the page prints no free allowance |
| 2026-10-30 | cloud.google.com/free/docs/free-cloud-features | the Google Cloud Free Tier excludes GPUs |
| 2026-10-30 | cloud.google.com/free/docs/free-cloud-features | while the billing account is a non-billable Free Trial account, GPUs cannot be added to VM instances |
| 2026-10-30 | cloud.google.com/products/compute/pricing/accelerator-optimi | an eight-H100 a3-highgpu-8g is billed USD 88.490000119 an hour on demand, below the a3-megagpu-8g at USD 93.400712807 |
| 2026-11-03 | cloud.google.com/free/docs/free-cloud-features | after the upgrade to a Paid billing account the remaining credit stays usable on products and services covered by the Free Trial, within 90 days of sign-up |
| 2026-11-03 | cloud.google.com/free/docs/free-cloud-features | the credit has its own exclusion list, which names Gemini API in AI Studio and managed-API partner models, and neither GPUs nor Windows Server images |
| 2026-11-03 | cloud.google.com/terms/free-trial | the Free Trial terms let an upgraded customer keep using unused credit for the Services, expiring 3 months after the Free Trial start |
| 2026-11-03 | cloud.google.com/free/docs/free-cloud-features | upgrading to a Paid billing account unlocks GPUs, which a non-billable Free Trial account cannot add |
| 2026-11-03 | groups.google.com/g/gce-discussion/c/JoOwDMtMgFk | a Google Cloud Support reply (2017) states the remaining free-trial credit is applied to GPU usage after the upgrade |
| 2026-11-03 | cloud.google.com/free/docs/free-cloud-features | the Free Trial signup, which precedes the upgrade, requires a credit card or other payment method |
More fields 5
| keep rate basis | no field in the source |
|---|---|
| billing model | unclassified |
| shape | {"vcpu": null,"ram": null,"disk": null,"gpu_class": "A100 80GB","gpu_count": 1,"note": "each accelerator-optimized type fixes its own ratio; a2-ultragpu-1g is 1 GPU, 12 vCPU, 170 GB, 275 GiB SSD"} |
| session cap | {"wall_clock_min": null,"idle_reap_s": null,"restart_required": false,"note": "no free allowance and no session cap; a GPU VM bills until it is stopped or deleted"} |
| free tier kind | none |
Unknown values 9
- chaining: a source was read and is silent on this field
- egress metered: egress metering was not published
- egress terms: a source was read and is silent on this field
- interface: the access interface was not established
- per object cap stated: a source was read and is silent on this field
- reliability note: nothing of this kind was observed
- retention notes: the field may be stated in a source that was not read
- retention stated: a source was read and is silent on this field
- tooling: a source was read and is silent on this field