Usage-based pricing models in Singapore: 2026 guide

Key takeaways:

• Per-unit, tiered, volume, and hybrid pricing can produce different bills for the same usage. The term “tiered” can also mean different things depending on the provider.

• Start by choosing the right usage metric, then decide how to price it. Even a well-designed pricing model won’t work if it measures the wrong thing.

• Airwallex Usage-Based Billing supports per-unit, volume, and graduated tiered pricing, so you can test different approaches without rebuilding your billing setup.

When choosing a usage-based pricing model in Singapore, start with how you want to charge customers for what they use.

The main options are per-unit, tiered, volume, and hybrid pricing. They may sound similar, but they can produce very different bills for the same level of usage.

This guide breaks down each model with examples, compares graduated and volume pricing, and explains how to choose a pricing model and usage metric.

For the billing requirements that come with usage-based pricing, including GST and InvoiceNow, see our guide to usage-based billing in Singapore.

Usage-based pricing models at a glance

The main difference between usage-based pricing models is how they calculate the final bill. Here’s a quick overview:

Revenue predictability refers to how easy it is to forecast income based on the pricing model alone. A fixed base fee or prepaid balance gives you some revenue upfront, while pure per-unit pricing depends entirely on how much customers use.

You can also combine models. For example, a hybrid plan can include a base fee and charge for extra usage using graduated tiered pricing.

Not sure which pricing model fits your business? Airwallex Usage-Based Billing supports per-unit, volume, and graduated tiered pricing, so you can choose a structure that suits your product and adjust your rates as your business grows.

Option 1: Per-unit pricing

Per-unit pricing charges the same rate for every unit a customer uses. It’s the simplest usage-based model and works as pay-as-you-go when there’s no minimum commitment.

When to use it: Per-unit pricing works well when each unit costs roughly the same to serve and customers want a clear breakdown of their charges.

The trade-off is that revenue depends entirely on usage, with no guaranteed minimum. Customers with high usage may also ask for lower rates.

If you want to offer discounts at higher usage levels, tiered or volume pricing may be a better fit. Adding a fixed base fee creates a hybrid model, which we cover later.

Per-unit pricing example

Imagine a Singapore document-processing API that charges S$10 per 1,000 API calls.

• 30,000 calls in a month: 30 × S$10 = S$300

• 8,000 calls the next month: 8 × S$10 = S$80 

The customer pays according to actual usage, so the bill rises or falls each month.

Option 2: Tiered (graduated) pricing

Graduated tiered pricing divides usage into bands, with a different rate for each band. As usage increases, the rate for additional units gets lower.

When to use it: Graduated tiered pricing works well when you want to reward higher usage while keeping rates fair for smaller customers. Each band is charged at its own rate, so reaching a lower-priced band doesn’t change the price of units already used.

Tiered pricing example

Using the same document-processing API, let’s say the rates are as follows:

In this example, here’s how the pricing is calculated:

• 8,000 calls: 8 × S$10 = S$80

• 30,000 calls: (10 × S$10) + 20 x S$8) = S$260

• 60,000 calls: (10 × S$10) + (40 x S$8) + (10 x S$6) = S$480

The key distinction is that each rate applies only to the calls within its band.

One thing to watch: “tiered” can also refer to plan packages such as Starter, Pro, and Enterprise. These group features together, rather than setting prices by usage. A plan can have both package tiers and graduated usage pricing.

Option 3: Volume pricing

Volume pricing uses a customer’s total usage to determine one rate, which then applies to every unit. It’s straightforward to calculate, but crossing a pricing threshold can sometimes make the total bill go down even as usage increases.

When to use it: Volume pricing can encourage customers to consolidate more of their usage with one provider. It also makes invoices relatively easy to check, since customers pay the same rate for every unit within a billing period.

The main trade-off is the sudden change in price when usage crosses a threshold.

Volume pricing example

Using the same illustrative rate card, the rate depends on the customer’s total monthly API calls:

• 8,000 calls: 8 × S$10 = S$80

• 30,000 calls: 30 × S$8 = S$240

• 60,000 calls: 60 × S$6 = S$360

Unlike graduated tiered pricing, the lower rate applies to every call once a customer reaches a new band.

With volume pricing, note that a pricing threshold can create a sharp drop in the total bill when a customer crosses into a cheaper band. Here’s an example:

For example, a customer making 11,000 calls pays less than one making 10,000, and a customer making 51,000 calls pays less than one making 50,000.

Customers may increase usage to reach a cheaper rate, but your revenue per customer can fall as a result.

Graduated vs volume pricing: how the bills compare

Graduated and volume pricing are sometimes both called “tiered pricing”, but they calculate charges differently. Here’s how the three models compare using the same API rates:

At 8,000 calls, all three models produce the same bill. As usage increases, the differences become more noticeable. At 60,000 calls, the volume-priced bill is S$120 lower than the graduated bill and S$240 lower than the per-unit bill.

The trade-off is how you want to reward higher usage. Per-unit pricing keeps the calculation simple, graduated pricing lowers the rate for additional units, and volume pricing applies the lower rate to all units once a threshold is reached.

That last approach can cause bills to drop when customers cross a threshold, so choose your pricing bands carefully.

Next, we’ll look at hybrid pricing, which combines a fixed base fee with usage charges to give you a more predictable revenue floor.

Option 4: Hybrid pricing (base fee plus usage)

Hybrid pricing combines a fixed recurring fee with additional charges based on usage. It gives you a more predictable revenue base while letting customers pay more as they use more.

When to use it: Hybrid pricing works well when you want recurring revenue without charging every customer the same total, regardless of usage. The trade-off is that light users may pay more per unit than they would under a pure per-unit model.

How it works: A hybrid plan can include four components:

• Base fee: The fixed amount customers pay each billing period.

• Included allowance: The usage covered by the base fee.

• Overage rate: The rate charged for usage beyond the allowance.

• Minimum commitment: An agreed minimum spend, which may be applied against usage charges.

Most straightforward plans use the first three components. Minimum commitments are more common in contracts with larger customers.

Hybrid pricing example

Using the same illustrative document-processing API, suppose the plan costs S$250 per month and includes 25,000 cals. Additional calls cost S$8 per 1,000.

• 8,000 calls: The customer pays the base fee of S$250, even though they use less than the included allowance.

• 30,000 calls: S$250 + (5 x S$8) = S$290

• 60,000 calls: S$250 + (35 x S$8) = S$530

For comparison, a customer making 8,000 calls would pay S$80 under per-unit pricing at S$10 per 1,000 calls. Hybrid pricing costs more for light users, but gives the business a more predictable revenue base.

How to set up a hybrid plan

• Set the included allowance close to typical customer usage so the base fee feels worthwhile.

• Price overage carefully so customers can see when upgrading to a larger plan makes sense.

• Decide whether unused allowance expires or rolls over, and explain the rules clearly.

• Offer usage alerts or spending limits to help customers avoid unexpected charges.

• Consider graduated rates for overage if you want to offer lower rates to customers with very high usage.

Pricing models for AI products: tokens, credits and outcomes

AI products can be harder to price because each request may cost a different amount to process.

A short question and a long document, for example, can require very different amounts of compute. The pricing model needs to account for these costs while making charges easy for customers to understand.

This section covers how to price AI products, while our guide to usage-based billing in Singapore explains how to manage the billing process.

If you’re wondering how to choose the right model, start with what your customers can easily understand and budget for.

Developers who know roughly how many tokens or API calls they need may prefer per-token or per-call pricing. Business buyers often find it easier to budget by task or through a monthly plan with included credits.

Per-task or outcome-based pricing can make costs easier for customers to predict, but you need to understand and control the cost of delivering each result. If tasks vary widely in complexity, a small number of expensive requests could eat into your margins.

How to choose a pricing model and usage metric

Choose your usage metric first, then decide how to price it. The metric determines what customers pay for, while the pricing model determines how much they pay.

If the metric doesn’t reflect the value customers get or the costs you incur, even a well-designed pricing model can fall short.

Start with the usage metric

Your usage metric is the unit you charge for, such as API calls, documents processed, messages sent, or gigabytes stored. Ideally, it should reflect both the value customers get from your product and the cost of serving them.

Avoid choosing a metric simply because it’s easy to track. Customers should be able to understand what they’re paying for and see how their usage affects the bill.

What makes a good usage metric? Consider these factors:

• Value: Can customers understand what they get from each unit?

• Measurability: Can your system reliably record every billable unit?

• Predictability: Can customers estimate their bills based on expected usage?

• Cost link: Does the metric reflect the resources needed to serve each unit?

Before launching, test the metric with a few existing customers. Ask them to estimate their bills using their own usage patterns, then compare their estimates with the actual charges. If they struggle to work out what they’ll pay, the metric or pricing structure may need revisiting.

Match the pricing model to your needs

The right pricing model depends on how customers use your product, what they value, and how predictable you want your revenue to be. Use the table below to narrow down the options that best fit your business:

Weigh the benefits against the risks

Usage-based pricing lowers the barrier to getting started because customers can begin with small amounts and pay more as their usage grows. It can also help your revenue grow alongside demand for your product.

The trade-off is less predictable revenue and more complex forecasting. Customers may also limit their usage to avoid higher bills, even when they could benefit from using your product more.

A fixed base fee or prepaid credits can help make revenue more predictable, while usage alerts and clear pricing rules help customers manage their spending.

Should you price in SGD, USD, or both?

If you sell to customers in Singapore and overseas, decide which currencies to quote your prices in and how to handle exchange rate changes. You also need to make sure your published prices comply with Singapore’s GST rules.

Choosing your rate card currency

Where possible, price in the currency your customers use.

If you sell in multiple currencies, publish a fixed rate card for each rather than converting prices in real time. This makes costs easier for customers to budget for.

If you incur expenses in USD, such as cloud hosting or compute, review your rates regularly to account for exchange rate movements and protect your margins.

Quoting prices with GST

If you are GST-registered, IRAS requires you to display and quote prices inclusive of GST to the public. This applies to websites, advertisements, and written or verbal quotes. You generally cannot display a public price as “S$100 + GST”.

The IRAS guidance does not specifically address per-unit rate cards or every business-to-business pricing scenario. If your pricing page is aimed at business customers, check with your tax adviser how the rules apply to your setup.

4 common usage-based pricing mistakes

Most usage pricing problems start with design choices that looked fine on a spreadsheet. Here are four to watch out for:

1. A metric customers cannot track. If they cannot estimate their usage, they cannot predict their bill.

2. Unchecked cliffs. Under volume pricing, a customer just below a threshold can pay more than one just above it.

3. Sudden changes for existing customers. Give notice early, and consider keeping loyal customers on their old terms for a set period.

4. No view of spend. Without alerts or caps, one busy month can become a dispute.

How to test your pricing before launch

To test your pricing before we launch it, run last quarter’s real usage through your new rate card, customer by customer. Compare each new bill with the old one, and look at the customers whose bills change most.

Then ask a small group of customers to estimate a bill from their own usage. If they cannot, rework the metric or the structure.

Offer usage-based pricing with Airwallex

Once you’ve chosen a pricing model, you need billing that can apply your rates accurately as customer usage changes.

Airwallex Usage-Based Billing supports per-unit, volume, graduated tiered, and hybrid pricing, combining recurring fees and usage charges in one subscription where needed.

Your pricing doesn’t have to stay fixed after launch. With Airwallex, you can adjust prices or usage tiers as your business evolves, and update your pricing without rebuilding your billing setup from scratch.

This saves your team from repeatedly changing billing logic in code whenever you want to test a new pricing structure.

You can sign up for an Airwallex account fully online; there are no setup fees or monthly fees on the free Explore plan.

Frequently asked questions (FAQs)

What is the difference between tiered and volume pricing?

In graduated tiered pricing, each band of usage is charged at its own rate. In volume pricing, total usage sets one rate that applies to every unit. In our illustrative example, 30,000 API calls cost S$260 under graduated pricing and S$240 under volume pricing. Volume pricing can also create cliffs, where a slightly larger customer pays less than a smaller one.

Is per-unit pricing the same as pay-as-you-go?

Not exactly. Per-unit pricing is a structure: one fixed rate for every unit. Pay-as-you-go describes how customers pay, which is for what they use, with no minimum commitment. Most pay-as-you-go plans use per-unit pricing, but you can also use per-unit pricing inside a plan that has a base fee.

What is hybrid pricing in SaaS?

Hybrid pricing combines a fixed recurring fee with usage charges. The base fee usually includes an allowance, and customers pay an overage rate for units beyond it. This gives you a predictable revenue floor while heavier users still pay more. Airwallex lets you combine recurring fees with usage-based charges in one subscription.

Which usage-based pricing model is best for a Singapore SaaS business?

There is no single best model. Choose per-unit pricing if every unit costs about the same to serve, graduated tiered pricing if you want to reward growth, and hybrid pricing if buyers need a predictable monthly cost. Start with your usage metric, then test each model on last quarter’s real usage before you decide.

How do I choose a usage metric for pricing?

Choose a metric that customers see as value, that you can measure reliably, and that tracks your cost to serve. Customers should also be able to estimate their own bill from their own plans. Test the metric with a few existing customers before you launch.

Does usage-based pricing work for AI products?

Yes, because the cost of serving each request varies. Common structures are per token or call, prepaid credits, per task, and a subscription with credits. Match the structure to what your buyer can predict: developers often prefer per-token pricing, while business buyers usually find credits or hybrid plans easier to buy. Airwallex Usage-Based Billing supports per-unit, volume, and graduated tiered pricing for AI and SaaS products.

Sources:

1. https://www.iras.gov.sg/taxes/goods-services-tax-(gst)/basics-of-gst/invoicing-price-display-and-record-keeping/displaying-and-quoting-prices