Vultr Pricing 2026: What Does the Cheapest $2.50 Plan Run for AI Bots?

Vultr pricing 2026 table for AI bots

🇻🇳 Đọc bản tiếng Việt

Vultr pricing currently starts at $2.50 a month for an IPv6-only plan, while the $5 plan gives you 1 vCPU, 1 GB of RAM and 25 GB of SSD in 31 locations, including Singapore. If you want to run an AI model on the server itself, an NVIDIA L40S GPU costs $1.671 an hour. The Vultr pricing table below comes from Vultr’s public plan list on November 4, 2026. We have not tested every plan ourselves.

Try Vultr

This is an affiliate link: it costs you nothing extra, and we may earn a commission.

Vultr pricing 2026 table for AI bots

Key Takeaways

  • Cheapest paid plan: $2.50 a month, 512 MB RAM, 10 GB SSD, IPv6 only, and only in New Jersey and Atlanta.
  • Most practical plan: $5 a month, 1 vCPU, 1 GB RAM, 25 GB SSD, 1 TB bandwidth, in 31 locations.
  • Asia: the $5 plan is offered in Singapore, Tokyo, Osaka and Seoul.
  • Hourly billing: the $5 plan is $0.007 an hour, capped at 672 hours a month.
  • GPU: NVIDIA L40S with 48 GB VRAM at $1.671 an hour.
  • Watch out: a stopped server is still billed. Only destroying it ends the charges.

What Is a VPS, and Why Does an AI Bot Need One?

A VPS (virtual private server) is a slice of a server in a data center that runs around the clock with its own operating system. Telegram bots, Discord bots and automations usually need to stay online, so a VPS beats leaving your home computer on all day.

Most AI bots don’t do the thinking on the server. The bot receives a message, sends the question to the OpenAI, Claude or Gemini API, and returns the answer. The heavy lifting happens at the model provider, so your server only needs to be small and reliable. That is why the low end of Vultr pricing is worth a close look.

Vultr Pricing Table for Cloud Compute

Vultr pricing as of November 4, 2026, from Vultr’s public plan list (api.vultr.com/v2/plans). These are the regular shared-CPU Cloud Compute plans.

PlanvCPURAMSSDBandwidthMonthlyHourlyLocations
IPv6 only1512 MB10 GB0.5 TB$2.50$0.0032 (New Jersey, Atlanta)
512 MB1512 MB10 GB0.5 TB$3.50$0.0051 (New Jersey)
1 GB11 GB25 GB1 TB$5$0.00731
2 GB12 GB55 GB2 TB$10$0.01431
2 vCPU22 GB65 GB3 TB$15$0.02131
4 GB24 GB80 GB3 TB$20$0.02731
8 GB48 GB160 GB4 TB$40$0.05531

Vultr pricing is the same in most locations. The exception we found is São Paulo, Brazil, where the 1 GB plan is $7.50 and the 2 GB plan is $15.

Two other lines cost a little more. The cheapest High Frequency plan is $6 for 1 vCPU, 1 GB RAM, 32 GB SSD and 1 TB bandwidth. The cheapest High Performance plan is also $6, in AMD and Intel versions, with 1 GB RAM, 25 GB SSD and 2 TB bandwidth.

What Does “IPv6 Only” Mean on the $2.50 Plan?

IPv6 is the newer kind of internet address. Many home networks and online services still rely on IPv4. An IPv6-only server has no public IPv4 address, so you can only reach it from a network with IPv6, and your bot can only call services that support IPv6.

The plan is also limited to New Jersey and Atlanta. A search result pointing to an older Vultr FAQ says the $2.50 plan is meant for testing and limited to two instances per account. That FAQ blocks bots, so we could not read it directly. Check the terms when you deploy.

In short, within Vultr pricing the $2.50 plan suits learning Linux or experimenting. For a bot you rely on, the $5 plan saves a lot of hassle.

Does Vultr Have a Free Plan?

The public plan list includes a $0 entry (vc2-1c-0.5gb-free) with 1 vCPU, 512 MB RAM and 10 GB SSD in Seattle, Frankfurt and Miami. We could not read Vultr’s pricing or promotions pages (both block bots), so we don’t know the conditions for that plan or whether a trial credit is on offer. Don’t plan a long-running bot on the $0 plan until you have seen its terms in the control panel.

What Can the $5 Singapore Plan Do for AI?

With 1 vCPU and 1 GB of RAM, the $5 plan covers light work: a chat bot that calls a model API, a script that pulls news every hour, or a webhook that receives data and passes it on. Singapore is the closest location in the list to Southeast Asia, so latency from the region will be lower than from a US server.

This plan is not for running a language model on the server itself; see the GPU section below for that. If you plan to run n8n with several workflows at once, the 2 GB ($10) or 4 GB ($20) plan is the sensible next step.

Vultr GPU Prices for Running Your Own Model

The cheapest GPU plan we found in the list is the NVIDIA L40S: 16 vCPUs, about 180 GB of RAM and 48 GB of VRAM for $1.671 an hour. Run it for a whole month and Vultr lists $1,122.91, because GPUs are billed for up to 730 hours a month. The two-GPU version (96 GB VRAM) is $3.342 an hour.

For individuals, the sensible approach is to rent by the hour and destroy the server when you are done, rather than leave it running all month.

How Does Vultr Pricing Work by the Hour?

According to Vultr’s support docs:

  1. The minimum billing unit is one hour.
  2. Regular servers are billed up to 672 hours a month. GPUs are billed up to 730 hours.
  3. Charges start when the server is deployed, whether it is on or off.
  4. Stopping a server does not stop billing. You have to destroy it.
  5. Invoices are issued on the 1st of each month for the previous period.

To sign up, create an account, pick Cloud Compute, choose Singapore and the 1 GB plan: Create a Vultr account.

For an overall verdict on the service, read our Vultr review. If you plan to automate with local AI, see our n8n and Ollama email workflow and our Ollama prompt templates.

Is Vultr Worth It for Running an AI Bot?

Pros: transparent Vultr pricing, hourly billing so three days on the $5 plan costs under a dollar, the $5 plan in 31 locations including Singapore, and GPUs you can rent by the hour.

Cons: the $2.50 plan looks great on paper but is IPv6 only and US only. Stopped servers are still billed, which is easy to forget. GPUs get expensive if left running. We could not read the promotions page, so we cannot confirm a trial credit.

Our take: if you just need a bot that calls an API around the clock, the $5 Singapore plan is a sensible start, and you know the cost up front. Skip it if your bot only runs a few hours a week; your own computer is enough. And don’t rent a GPU by the month just to try a model. Rent it by the hour and destroy it as soon as you are done.

Vultr Pricing FAQ

What is the cheapest Vultr plan?
The cheapest paid plan is $2.50 a month and IPv6 only. The cheapest plan with IPv4 that is widely available is $5 a month.

Does Vultr have servers in Vietnam?
Not in the list. Vultr pricing covers 31 locations for the $5 plan, and the closest ones are Singapore, then Tokyo, Osaka and Seoul.

Am I charged for a stopped Vultr server?
Yes. Vultr’s docs say stopped servers keep accruing hourly charges. Only destroying the server ends billing.

Can a 1 GB plan run an AI model?
It is not a good fit. It suits bots that call an API. To run a model yourself, you need more memory or a GPU.

How much is a Vultr GPU?
The NVIDIA L40S with 48 GB VRAM is $1.671 an hour, up to $1,122.91 a month.

Sources

Last updated: November 4, 2026

Trạm Công Nghệ AI
Logo
Compare items
  • Total (0)
Compare
0