# Vultr Cloud Compute

Official pricing: https://www.vultr.com/pricing/  
Category: hyperscaler · Isolation: vm

## Pricing regimes (raw)

- **Cloud Compute Regular (vc2, shared vCPU), hourly with 672 h monthly cap** (sizes): vc2-1c-1gb (Cloud Compute Regular, shared) 1 vCPU/1 GiB $0.0074/h; vc2-1c-2gb (Cloud Compute Regular, shared) 1 vCPU/2 GiB $0.0149/h; vc2-2c-2gb (Cloud Compute Regular, shared) 2 vCPU/2 GiB $0.0223/h; vc2-2c-4gb (Cloud Compute Regular, shared) 2 vCPU/4 GiB $0.0298/h; vc2-4c-8gb (Cloud Compute Regular, shared) 4 vCPU/8 GiB $0.0595/h; vc2-6c-16gb (Cloud Compute Regular, shared) 6 vCPU/16 GiB $0.119/h
- **Cloud Compute High Performance (vhp, AMD EPYC or Intel Xeon, NVMe, shared vCPU), hourly with 672 h cap** (sizes): vhp-1c-1gb-amd (High Performance AMD, shared) 1 vCPU/1 GiB $0.0089/h; vhp-1c-2gb-amd (High Performance AMD, shared) 1 vCPU/2 GiB $0.0179/h; vhp-2c-2gb-amd (High Performance AMD, shared) 2 vCPU/2 GiB $0.0268/h; vhp-2c-4gb-amd (High Performance AMD, shared) 2 vCPU/4 GiB $0.0357/h; vhp-4c-8gb-amd (High Performance AMD, shared) 4 vCPU/8 GiB $0.0714/h; vhp-4c-12gb-amd (High Performance AMD, shared) 4 vCPU/12 GiB $0.1071/h
- **Cloud Compute High Frequency (vhf, 3 GHz+ Intel, NVMe, shared vCPU), hourly with 672 h cap** (sizes): vhf-1c-1gb (High Frequency Intel 3GHz+, shared) 1 vCPU/1 GiB $0.0089/h; vhf-1c-2gb (High Frequency Intel 3GHz+, shared) 1 vCPU/2 GiB $0.0179/h; vhf-2c-2gb (High Frequency Intel 3GHz+, shared) 2 vCPU/2 GiB $0.0268/h; vhf-2c-4gb (High Frequency Intel 3GHz+, shared) 2 vCPU/4 GiB $0.0357/h; vhf-3c-8gb (High Frequency Intel 3GHz+, shared) 3 vCPU/8 GiB $0.0714/h; vhf-4c-16gb (High Frequency Intel 3GHz+, shared) 4 vCPU/16 GiB $0.1429/h
- **Optimized Cloud Compute (voc, dedicated AMD EPYC vCPU: General / CPU / Memory / Storage Optimized), hourly with 672 h cap** (sizes): voc-c-1c-2gb-25s-amd (Optimized Cloud Compute CPU Optimized, dedicated) 1 vCPU/2 GiB $0.0417/h; voc-g-1c-4gb-30s-amd (Optimized Cloud Compute General Purpose, dedicated) 1 vCPU/4 GiB $0.0446/h; voc-m-1c-8gb-50s-amd (Optimized Cloud Compute Memory Optimized, dedicated) 1 vCPU/8 GiB $0.0595/h; voc-c-2c-4gb-50s-amd (Optimized Cloud Compute CPU Optimized, dedicated) 2 vCPU/4 GiB $0.0595/h; voc-g-2c-8gb-50s-amd (Optimized Cloud Compute General Purpose, dedicated) 2 vCPU/8 GiB $0.0893/h; voc-c-2c-4gb-75s-amd (Optimized Cloud Compute CPU Optimized, dedicated) 2 vCPU/4 GiB $0.067/h
- **VX1 Cloud Compute (dedicated AMD vCPU, hourly, NOT capped: billed on actual hours)** (sizes): vx1-g-2c-8g (VX1 General Purpose, dedicated, boot from block storage (billed separately)) 2 vCPU/8 GiB $0.06/h; vx1-m-2c-16g (VX1 Memory Optimized, dedicated, boot from block storage (billed separately)) 2 vCPU/16 GiB $0.08/h; vx1-g-4c-16g (VX1 General Purpose, dedicated, boot from block storage (billed separately)) 4 vCPU/16 GiB $0.12/h; vx1-m-4c-32gb (VX1 Memory Optimized, dedicated, boot from block storage (billed separately)) 4 vCPU/32 GiB $0.16/h; vx1-g-8c-32g (VX1 General Purpose, dedicated, boot from block storage (billed separately)) 8 vCPU/32 GiB $0.24/h; vx1-m-8c-64gb (VX1 Memory Optimized, dedicated, boot from block storage (billed separately)) 8 vCPU/64 GiB $0.32/h
- **Bare Metal (single-tenant, hourly with 672 h cap)** (sizes): vbm-4c-32gb (Bare Metal Intel E3-1270, 4c/8t, 2x240 GB SSD) 8 vCPU/32 GiB $0.1786/h; vbm-8c-132gb (Bare Metal Intel E-2288G, 8c/16t, 2x1920 GB NVMe) 16 vCPU/128 GiB $0.5208/h; vbm-6c-32gb (Bare Metal Intel E-2286G, 6c/12t, 2x960 GB SSD) 12 vCPU/32 GiB $0.2753/h; vbm-24c-256gb-amd (Bare Metal AMD EPYC 7443P, 24c/48t, 2x480 GB SSD) 48 vCPU/256 GiB $1.0789/h; vbm-8c-132gb-v2 (Bare Metal Intel E-2388G, 8c/16t, 2x1920 GB NVMe) 16 vCPU/128 GiB $0.5208/h; vbm-24c-384gb-amd (Bare Metal AMD EPYC 9254, 24c/48t, 2x480 GB SSD) 48 vCPU/384 GiB $1.2277/h

## Features

Yes: snapshots, fork/clone, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, Docker inside, root, systemd, SSH, public IPv4, custom domains, raw TCP inbound, open internet, static egress IP, private networking, SSO, GPU, Windows, audit logs, EU data residency, snapshots on demand, inbound access rules, firewall inside (nftables), extra volumes, shared volumes

No: memory snapshots, pause/resume, idle auto-stop, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, preinstalled agents, HTTPS preview URLs, egress allowlist, self-hosting, BYOC, open source, arm64, macOS, GPU desktop, HTTP method/path egress rules, wake on request, memory fork, automatic snapshots, agent harness API, secret proxy, secret proxy for any API

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

## Caveats

- DIY infrastructure rather than an agent-managed sandbox: guest installation, HTTPS/domain setup, abuse controls and operations are customer responsibilities. Linux guest administrative capabilities are not preinstalled agent features. 30% guest CPU utilization does not reduce allocated-compute charges. Unknowns remain null; a partial subtotal is not a complete invoice.
- Hourly price discrepancy: vultr.com/pricing (Wayback 2026-09-26) and the billing doc imply hour = monthly/672 with a 672 h cap ($40/mo -> $0.060/h), but the public /v2/plans API returns hourly_cost = monthly/730 (e.g. vc2-4c-8gb 0.055). The card uses monthly/672 (the documented billing rule); if Vultr actually bills the API rate, part-time costs are ~8% lower.
- Minimum billing unit is 1 hour (every started hour billed). Stopped instances and stopped bare metal bill in full until destroyed.
- GPU instances bill on 730 h (actual hours) since 2024-10-01; VX1 bills actual hours with no 672 h cap.
- Egress: 2 TB free per account per month plus each instance's allowance accrued hourly (monthly allowance/672 per hour), pooled; overage $0.01/GB; inbound free. egress_free_gib here = 2,048 account + 4,096 of a vc2-4c-8gb (per-instance, prorated) - not an account constant.
- Snapshots $0.05/GB-month on compressed size, usable in any region; deploy new instances from a snapshot to clone (fs only, no memory). No snapshots for bare metal.
- One public IPv4 per instance is included; Reserved IPs cost $3/month; additional instance IPv4 price not captured.
- Pricing may vary by location (e.g. Sao Paulo is 1.5x on many plans in the API location_cost field).
- Nested virtualization: documented for VX1 ('support for virtualization') and inherent on bare metal; not documented for vc2/vhp/vhf/voc (null).
- Vultr Firewall filters inbound only (egress filtering must be done in the guest).
- No published commitment/reserved discounts; free-tier plan vc2-1c-0.5gb-free (1 vCPU/0.5 GB, $0) exists in SEA/FRA/MIA per API; promo-code credits not captured.
- Default account instance limit not published; 50 concurrent instances likely needs a limit increase.
- vultr.com is Cloudflare-protected; website figures come from the Wayback snapshot of 2026-09-26 and docs.vultr.com; plan specs from the unauthenticated API on 2026-09-28.
- Boot / provisioning time is not published for vc2, vhp, vhf, voc or VX1 ("ready for use in seconds" is the only VX1 claim), so boot_overhead_s stays null; with 1-hour minimum billing, boot time sits inside the first billed hour anyway. A new instance also needs OS setup (DIY VM, no sandbox API).
- Automatic backups add 20% to the instance price; DDoS protection $10/mo per instance; Reserved IPs $3/mo ($0.004/h); promo-code credits expire and require a linked card/PayPal (one promo per user).
- Free Tier Program: application-only, randomized/weighted admission, limited regions; API plan vc2-1c-0.5gb-free (1 vCPU / 0.5 GB / 10 GB, $0) is deployable in sea/fra/mia. Not modelled as free credit.

## How this provider charges


DIY infrastructure rather than an agent-managed sandbox: guest installation, HTTPS/domain setup, abuse controls and operations are customer responsibilities. Linux guest administrative capabilities are not preinstalled agent features. 30% guest CPU utilization does not reduce allocated-compute charges. Unknowns remain null; a partial subtotal is not a complete invoice.

Prices below are public list prices unless a row explicitly says promotion/commitment. USD; taxes excluded unless noted.

| Regime | When / billing | Captured prices | Source |
|---|---|---|---|
| Cloud Compute Regular (vc 2, shared vCPU), hourly with 672 h monthly cap | On demand; Hourly billing, minimum 1 h, capped at the monthly price after 672 h/month. hour = monthly/672 (matches vultr.com/pricing, e.g. $40/mo = $0.060/h). The public API's hourly_cost field shows monthly/730 instead (e.g. 0.055) - see caveats. vc 2 = previous-gen Intel, local SSD. Billed while stopped until destroyed. | vc 2-4c-8gb (Cloud Compute Regular, shared): $0.0595/h; cap 40 | https://www.vultr.com/pricing/ |
| Cloud Compute High Performance (vhp, AMD EPYC or Intel Xeon, NVMe, shared vCPU), hourly with 672 h cap | On demand; AMD and Intel variants list identical prices; AMD list used. vhp-4c-8gb $48/mo = $0.0714/h. Max 12 vCPU / 24 GB. | vhp-4c-8gb-amd (High Performance AMD, shared): $0.0714/h; cap 48 | https://www.vultr.com/pricing/ |
| Cloud Compute High Frequency (vhf, 3 GHz+ Intel, NVMe, shared vCPU), hourly with 672 h cap | On demand; No 4 vCPU/8 GB shape (vhf-3c-8gb, vhf-4c-16gb $96/mo). | vhf-4c-16gb (High Frequency Intel 3GHz+, shared): $0.1429/h; cap 96 | https://www.vultr.com/pricing/ |
| Optimized Cloud Compute (voc, dedicated AMD EPYC vCPU: General / CPU / Memory / Storage Optimized), hourly with 672 h cap | On demand; Dedicated vCPU. CPU Optimized voc-c-4c-8gb-75s $80/mo = $0.119/h; General Purpose voc-g-4c-16gb $120/mo = $0.1786/h. | voc-c-4c-8gb-75s-amd (Optimized Cloud Compute CPU Optimized, dedicated): $0.119/h; cap 80 | https://www.vultr.com/pricing/ |
| VX1 Cloud Compute (dedicated AMD vCPU, hourly, NOT capped: billed on actual hours) | On demand; VX1 (docs published 2025-10-17): dedicated CPU with virtualization support, 'instant provisioning in under 15 seconds' (vendor claim), billed on actual hours each month and NOT capped at 672 h (month_cap null). Plans without '-###s' boot from high-perf NVMe block storage billed separately (~$0.10/GB-month, disk_gib 0 here so the engine adds block storage); '-###s' plans include local NVMe. No automatic backups, no Windows. 9-11 locations (NA/EU/Sydney/Tokyo). | vx 1-g-4c-16g (VX1 General Purpose, dedicated, boot from block storage (billed separately)): $0.12/h; cap None | https://www.vultr.com/pricing/ |
| Bare Metal (single-tenant, hourly with 672 h cap) | On demand; Single-tenant physical servers, vcpu = hardware threads. Hourly at monthly/672 (e.g. vbm-4c-32gb E3-1270 $120/mo = $0.179/h), minimum 1 h, capped at the monthly price; billed while stopped. No snapshots on bare metal. Provisioning takes longer than VMs (not quantified). GPU bare metal (H100/B200/MI300X/MI325X/MI355X...) is mostly preemptible/contract-only in the API and not modelled. | vbm-4c-32gb (Bare Metal Intel E3-1270, 4c/8t, 2x 240 GB SSD): $0.1786/h; cap 120 | https://www.vultr.com/pricing/ |

## Billing details
- **granularity_seconds**: 3600
- **minimum_billed_seconds**: 3600
- **setup_fee_usd**: 0
- **cpu**: Cloud Compute regular/high-performance/high-frequency shared; Optimized dedicated; VX1 dedicated; bare metal physical server.
- **backup**: Automatic backups add 20% of instance cost; compressed VM snapshots $0.05/GB-month; not available on bare metal.
- **api**: True
- **cli**: True
- **ipv 6**: True
- **ingress_usd**: 0
- **minimum_term**: One started hour; destroy to stop billing.
- **provisioning**: No independently measured deployment timing; VM and physical provisioning differ.

## Gotchas
- DIY infrastructure rather than an agent-managed sandbox: guest installation, HTTPS/domain setup, abuse controls and operations are customer responsibilities. Linux guest administrative capabilities are not preinstalled agent features. 30% guest CPU utilization does not reduce allocated-compute charges. Unknowns remain null; a partial subtotal is not a complete invoice.
- DIY infrastructure rather than an agent-managed sandbox: guest installation, HTTPS/domain setup, abuse controls and operations are customer responsibilities. Linux guest administrative capabilities are not preinstalled agent features. 30% guest CPU utilization does not reduce allocated-compute charges. Unknowns remain null; a partial subtotal is not a complete invoice.
- DIY infrastructure rather than an agent-managed sandbox: guest installation, HTTPS/domain setup, abuse controls and operations are customer responsibilities. Linux guest administrative capabilities are not preinstalled agent features. 30% guest CPU utilization does not reduce allocated-compute charges. Unknowns remain null; a partial subtotal is not a complete invoice.
- Hourly price discrepancy: vultr.com/pricing (Wayback 2026-09-26) and the billing doc imply hour = monthly/672 with a 672 h cap ($40/mo -> $0.060/h), but the public /v 2/plans API returns hourly_cost = monthly/730 (e.g. vc 2-4c-8gb 0.055). The card uses monthly/672 (the documented billing rule); if Vultr actually bills the API rate, part-time costs are ~8% lower.
- Minimum billing unit is 1 hour (every started hour billed). Stopped instances and stopped bare metal bill in full until destroyed.
- GPU instances bill on 730 h (actual hours) since 2024-10-01; VX1 bills actual hours with no 672 h cap.
- Egress: 2 TB free per account per month plus each instance's allowance accrued hourly (monthly allowance/672 per hour), pooled; overage $0.01/GB; inbound free. egress_free_gib here = 2,048 account + 4,096 of a vc 2-4c-8gb (per-instance, prorated) - not an account constant.
- Snapshots $0.05/GB-month on compressed size, usable in any region; deploy new instances from a snapshot to clone (fs only, no memory). No snapshots for bare metal.
- One public IPv 4 per instance is included; Reserved IPs cost $3/month; additional instance IPv 4 price not captured.
- Pricing may vary by location (e.g. Sao Paulo is 1.5x on many plans in the API location_cost field).
- Nested virtualization: documented for VX1 ('support for virtualization') and inherent on bare metal; not documented for vc 2/vhp/vhf/voc (null).
- Vultr Firewall filters inbound only (egress filtering must be done in the guest).
- No published commitment/reserved discounts; free-tier plan vc 2-1c-0.5gb-free (1 vCPU/0.5 GB, $0) exists in SEA/FRA/MIA per API; promo-code credits not captured.
- Default account instance limit not published; 50 concurrent instances likely needs a limit increase.
- vultr.com is Cloudflare-protected; website figures come from the Wayback snapshot of 2026-09-26 and docs.vultr.com; plan specs from the unauthenticated API on 2026-09-28.

## Worked example
4 vCPU / 8 GiB; 50 concurrent × 8 h/day × 22 days = **8,800 guest-hours**; 30% CPU busy; **50 GiB-month total snapshots**, **100 GiB total egress**. 50 GiB retained state is total, not 50 GiB per VM. Compute-only subtotals below exclude unknown components and tax.

| Mode | Delete/recreate / active-window compute | 50 retained hosts, 730-hour month |
|---|---:|---:|
| Cloud Compute Regular (vc 2, shared vCPU), hourly with 672 h monthly cap | $523.60 | $2,000.00 |
| Cloud Compute High Performance (vhp, AMD EPYC or Intel Xeon, NVMe, shared vCPU), hourly with 672 h cap | $628.32 | $2,400.00 |
| Cloud Compute High Frequency (vhf, 3 GHz+ Intel, NVMe, shared vCPU), hourly with 672 h cap | $1,257.52 | $4,800.00 |
| Optimized Cloud Compute (voc, dedicated AMD EPYC vCPU: General / CPU / Memory / Storage Optimized), hourly with 672 h cap | $1,047.20 | $4,000.00 |
| VX1 Cloud Compute (dedicated AMD vCPU, hourly, NOT capped: billed on actual hours) | $1,056.00 | $4,380.00 |
| Bare Metal (single-tenant, hourly with 672 h cap) | $1,571.68 | $6,000.00 |

Snapshot subtotal: $2.5000; rate $0.05/GiB-month. Egress: see allowances, pooling/accrual and rate $0.01/GiB; not every mode has the same allowance.
Powered-off compute billed: True. A retained-host column is the always-running upper scenario when this is false; powered-off storage/IP still accrue.
Monthly caps attach to instance identities, not to a pool of replacement instances. The active-window column requires deleting/releasing billable compute between shifts for products that bill while off. Restoring from disk is not memory resume. The engine cannot perfectly model these lifecycles; see validation/estimator-gaps.md.

## Sources
- https://www.vultr.com/pricing/
- https://api.vultr.com/v2/plans
- https://api.vultr.com/v2/plans-metal
- https://api.vultr.com/v2/regions
- https://web.archive.org/web/20260926221221/https://www.vultr.com/pricing/
- https://docs.vultr.com/support/platform/billing/how-am-i-billed-for-my-servers
- https://docs.vultr.com/support/platform/billing/are-stopped-instances-still-billed-on-vultr
- https://docs.vultr.com/support/platform/billing/do-stopped-bare-metal-servers-incur-charges
- https://docs.vultr.com/support/platform/billing/how-are-bandwidth-caps-calculated
- https://docs.vultr.com/support/platform/billing/what-is-the-bandwidth-overage-rate
- https://docs.vultr.com/support/platform/billing/how-is-bandwidth-usage-calculated
- https://docs.vultr.com/support/platform/billing/does-vultr-charge-for-stored-snapshots
- https://docs.vultr.com/support/platform/billing/how-much-does-it-cost-to-enable-automatic-backups
- https://docs.vultr.com/support/platform/billing/how-are-gpu-products-billed-differently
- https://docs.vultr.com/support/platform/billing/is-pricing-the-same-in-all-data-center-locations
- https://docs.vultr.com/support/platform/billing/how-much-does-ddos-protection-cost
- https://docs.vultr.com/support/platform/billing/how-can-i-increase-my-account-limits
- https://docs.vultr.com/support/platform/billing/what-happens-when-i-upgrade-my-server
- https://docs.vultr.com/support/platform/billing/is-a-windows-license-included-in-the-monthly-price
- https://docs.vultr.com/support/products/network/are-reserved-ips-free
- https://docs.vultr.com/support/products/network/does-vultr-firewall-filter-outgoing-traffic-from-my-instance
- https://docs.vultr.com/support/products/compute/how-do-i-clone-or-make-a-copy-of-a-vultr-compute-instance
- https://docs.vultr.com/products/storage/snapshots/faq
- https://docs.vultr.com/products/storage/backups/faq
- https://docs.vultr.com/platform/billing/manage-account-limits
- https://docs.vultr.com/platform/billing/faq
- https://docs.vultr.com/vultr-vx1-cloud-compute
- https://docs.vultr.com/products/compute/instances/vx1-cloud-compute/management/stop
- https://docs.vultr.com/products/compute/instances/cloud-compute/networking/ipv4
- https://web.archive.org/web/20260621045803/https://www.vultr.com/servers/windows/