PRICING

2 M4 plans,
daily to quarterly billing

Starting at $20.2/day . Rent an entire dedicated physical machine, not a virtual machine. Choose a billing term to match your project timeline. All prices are in USD.

Both plans include SSH and VNC access. Start by choosing the memory and storage you need, then select a location and billing term. Add-ons are priced separately.

Choose your workload, then your term

One physical Mac, four ways to rent

Start with daily billing for short-term testing, or compare weekly, monthly, and quarterly terms for ongoing development. Prices below are for the full term, not a daily equivalent. Switching terms updates both plan cards, and the complete price list stays visible for easy comparison.

Showing daily pricing. See each plan card for prices on other terms.

Single-app builds and short regression tests

MetalBench 16

Chip
M4
Memory
16GB RAM
Storage
256GB SSD

$20.2/day

A good fit for building, testing, and archiving a single project. Budget for dependency caches, build artifacts, and project files—not just the repository—to avoid running short on disk space.

Daily
$20.2/day
Weekly
$54.6/week
Monthly
$101.2/month
Quarterly
$275.3/quarter

Available locations: Singapore / Tokyo / Seoul / Hong Kong

Available Rent MetalBench 16

Desktop debugging and multitasking

MetalBench 24

Chip
M4
Memory
24GB RAM
Storage
512GB SSD

$39.2/day

Well suited to multi-target projects, parallel simulators, and VNC desktop debugging. More memory and storage than the entry plan support more complex development environments, but test your project to determine the right concurrency level.

Daily
$39.2/day
Weekly
$105.8/week
Monthly
$195.9/month
Quarterly
$532.8/quarter

Available locations: Singapore / Tokyo / Seoul / Hong Kong

Available Rent MetalBench 24

Availability labels indicate that these configurations are generally orderable in the catalog. Check the console for real-time availability when placing your order. The buttons above only change the displayed prices; confirm your billing term in the order configuration.

Compare both plans side by side

Specs and full-term pricing for both M4 plans
Plan detailsMetalBench 16MetalBench 24
ChipM4M4
Memory16GB RAM24GB RAM
Storage256GB SSD512GB SSD
Daily$20.2/day$39.2/day
Weekly$54.6/week$105.8/week
Monthly$101.2/month$195.9/month
Quarterly$275.3/quarter$532.8/quarter
Resource allocationEntire dedicated physical machineEntire dedicated physical machine
Remote accessSSH command line and VNC desktopSSH command line and VNC desktop
Storage and connectivity, priced separately

Add storage when you need it—no hidden add-on costs in the starting price

Plan prices exclude the add-ons below. Check actual usage across source code, toolchains, simulators, caches, and build artifacts before deciding whether to upgrade. More SSD storage adds capacity, not unified memory, and is not a substitute for a higher-memory plan.

Four-term pricing for SSD and Thunderbolt 5 connectivity add-ons
Add-onDailyWeeklyMonthlyQuarterly
+1TB SSD$2.4/day$6.5/week$12.1/month$32.9/quarter
+2TB SSD$4.8/day$13/week$24.2/month$65.8/quarter
Thunderbolt 5 connection (per machine)$1.1/day$3/week$5.5/month$15/quarter

First, identify whether you need capacity or performance

If build caches need frequent clearing or model files won't fit, check disk usage first. If simulators and compiler processes are competing for memory, check memory pressure. More storage won't solve a memory bottleneck, and it doesn't guarantee faster builds.

Thunderbolt 5 is priced per machine

Thunderbolt 5 connectivity supports multi-machine workflows; it does not automatically distribute your application across machines. Before ordering multiple machines, share how many you need, how data will move, and how tasks will be split so you can confirm your app and storage workflow can make use of the connection.

To estimate your order, use the plan and add-on prices for the same billing term. Confirm the selected quantity, configuration, term, and total due at checkout.

Choose based on your first test run

Start with the plan that fits your workload

Recommendations are not performance guarantees. Run a complete test with your actual repository, dependencies, and dataset. Track build times, memory pressure, and peak storage use before increasing concurrency or extending your rental. Getting your workload running comes before committing to a longer term.

Plan and billing-term recommendations for light builds, everyday development, and MLX experiments
WorkloadRecommended planRecommended termWhy this plan and what to test
Light buildsMetalBench 16
M4 · 16GB · 256GB
Test daily first; compare weekly for ongoing tasksA fit for single-app CI and simulator regression tests. Run one clean build and one cached build, then check build times and disk growth for both.
Everyday developmentMetalBench 24
M4 · 24GB · 512GB
Monthly; consider quarterly for steady, long-running workloadsA fit for multi-target projects, parallel simulators, and VNC desktop debugging. Increase concurrency gradually and check responsiveness and test stability—don't estimate speed based on memory alone.
LLM inference experimentsMetalBench 24 +2TB SSD
M4 · 24GB · 512GB
Test models daily; consider quarterly for ongoing experimentsSuitable for MLX experiments with quantized 7B–13B models. Check the quantization format, context length, and runtime overhead first. The extra storage holds models and datasets; loading them still depends on memory requirements.

Before choosing a plan, gather these three data points

  1. 01Track peak resource usage

    Note memory pressure, disk usage, and the number of tasks running at once. Separate one-time dependency installs from everyday incremental builds so a single peak doesn't skew your estimate for the whole rental.

  2. 02Choose how you'll connect

    SSH may be enough if pipeline logs are all you need. Keep VNC if you need to inspect the UI, simulators, or failures firsthand. Desktop responsiveness also depends on your network route to the machine.

  3. 03Set an end date for the workload

    A short acceptance test, a release push, and ongoing development call for different rental terms. Include time for backups, exporting artifacts, and moving your environment—not just build time.

Look beyond the rental price

Rent, buy, or use a shared VM?

Compare delivery time, resource ownership, and the cost of changing your setup. A dedicated physical machine suits workloads that need a controlled toolchain, remote desktop, and a complete development environment. Local hardware still makes sense for offline work or workflows that rely on nearby peripherals.

Trade-offs between renting a dedicated cloud Mac, buying hardware, and using a shared VM
What to compareRent a dedicated physical MacBuy hardwareShared VM
Cost structureRent by the day; pay for your selected term and add-onsPay upfront for hardware, then cover depreciation, power, and idle timeMay have a lower starting price; check for resource contention and plan limits
Time to get startedTypically about 15 minutes; check progress in the consolePurchasing and delivery can take days, plus environment setupUsually available immediately, depending on the service
Resource exclusivityAn entire physical machine; CPU, memory, and storage aren't shared with other tenantsYour own physical device, with access managed by youShares a host with other users; isolation depends on the service
Upgrade flexibilityRequest a plan change or more storage; confirm migration steps and costs firstChanging the configuration may require replacing or adding hardwareOptions are limited by the host and service plan
ConnectivityConnect over SSH or VNC; a working internet connection is requiredUse locally; configure remote access yourself if neededConnect over the internet; check desktop and device support separately
Best suited toRelease work, remote development, isolated build environments, and model experimentsA permanent desk setup, offline work, and workflows with local peripheralsWorkloads that suit virtualization and can tolerate sharing a host

Don't decide based only on a single build time. Factor in environment setup, debugging failures, exporting artifacts, and idle time. This table compares usage models; it doesn't make up prices or benchmark scores for other services.

Two plans, four locations

All 8 plan-and-location combinations are listed

Choose from four Asia-Pacific locations: Singapore, Tokyo, Seoul, and Hong Kong. Both plans are available in every location. Compare actual network connectivity, not just distance—desktop performance can vary across providers within the same city.

Catalog availability of both plans across four locations; check the console for real-time availability
PlanSingaporeTokyoSeoulHong Kong
MetalBench 16AvailableAvailableAvailableAvailable
MetalBench 24AvailableAvailableAvailableAvailable

This is the current product catalog, not a timestamped inventory feed. Check the console for real-time availability. Nodes run year-round, 365 days a year. For SSH builds, prioritize connection stability; for VNC, test interactive latency too.

Check the details before you pay

Payments, renewals, and data handling

Plan cards show catalog prices; the order confirmation page lists the charges for your selected configuration. Before paying, check the plan, location, term, add-ons, and quantity. For issues with an existing machine, submit a support ticket through the console so it can be linked to your order and instance.

Which payment methods and currency do you support?

We accept USDT-TRC20 and Visa, Mastercard, or Amex credit cards via Stripe. All prices are in USD. Available payment gateways are determined by the backend at checkout. If paying with USDT, verify the network, order amount, and payment details. Transfers on other networks won't count as the same payment.

How should I compare daily, weekly, monthly, and quarterly terms?

Choose based on how long you'll need the machine, not just how long commands take to run. Daily billing works for short tests; for ongoing development, compare the full weekly, monthly, and quarterly prices. The price list shows each term's actual price—not a daily equivalent—and doesn't default to a long-term commitment. Before renewing, check your remaining tasks and the expiry date.

Are extra storage and Thunderbolt 5 connectivity included in the plan price?

No. +1TB SSD, +2TB SSD, and Thunderbolt 5 connectivity are separate add-ons. Thunderbolt 5 is priced per machine. Compare catalog prices for the same billing term and check the quantity you need. Don't add a daily add-on price to a monthly plan price and treat that as the order total; check the full breakdown at checkout.

Why does an order still need confirmation after I send USDT?

A transfer record and an order confirmation are different statuses. Blockchain confirmations, payment gateway recognition, and order updates happen in sequence; processing time depends on the transaction, so there is no guaranteed arrival time in minutes. Check your order status in the console after paying. If you need help, include your order number, transaction hash, network, and amount in a support ticket. Don't send another payment to try to speed things up.

Where do I renew, and how can I avoid an interruption when my rental expires?

Check your instance expiry date and billing details in the console, and renew before the expiry date. Schedule long tests, artifact exports, and backups within your rental term rather than relying on reminders. After submitting payment, check the billing page and the instance's new expiry date. If the status hasn't updated, verify it through a support ticket; initiating a payment doesn't mean the renewal is complete.

Can I upgrade my plan or add SSD storage if my workload grows?

You can request a plan change or storage expansion. Include your current instance, target configuration, remaining rental term, and any data you need to keep. Support will confirm the price difference, steps, and migration plan. A plan change doesn't migrate your app automatically, and a machine restart doesn't mean the upgrade is complete. Back up data on the working drive before making changes, and proceed once the plan is confirmed.

How long is my data kept after the rental expires?

Data retention and deletion timelines are governed by the service terms for your order; this page doesn't set a standard retention period. If you need to confirm the timeline before ordering, contact us by email. For an existing instance, ask about the specific arrangement in a support ticket. Treat the expiry date as your deadline for backups and exports—don't rely on post-expiry retention or wait until your instance expires to start migrating.

What billing details do I need for team expense claims?

Check your order and billing records in the console, then submit a support ticket with the billing name, billing entity, order number, and required document fields. Support will confirm which billing documents are available. USD payment records and tax documentation requirements vary by location. Clarify your team's finance requirements before renting; a payment record isn't necessarily a tax invoice in a specific format.

For pre-sales questions, email us at support@metalvnc.com; for existing orders, log in to the console and submit a support ticket. Include your order number and steps to reproduce the issue. Never send passwords, private keys, or full payment card details.

Validate your choice with a real build

Found the right setup?
Rent a machine and start testing

Choose the M4 plan that fits your workload, then confirm the term, location, and add-ons at checkout. Not sure? Tell us about your project size, concurrency needs, and expected rental period.