SaaS Pricing Models 101
A complete guide to the SaaS pricing models, pricing strategies, and how to choose — and change — the right one for your product.
What is a SaaS pricing model?
A SaaS pricing model is just the structure behind how you charge — flat fee, per seat, per unit of usage, whatever shape it takes. That's a different question from pricing strategy, which is deciding how you land on the actual number: what value you're charging for, and why. You need both — the model is the structure, the strategy fills it in.
SaaS pricing also works differently from the perpetual-license model it replaced. A perpetual license is a one-time purchase — the customer owns the software outright, and the vendor relationship largely ends at the sale. SaaS flips that: the vendor hosts and maintains the product, the customer pays on an ongoing basis, and the relationship — and the revenue — continues for as long as the subscription does.
For the deeper walkthrough of both model and strategy, see the complete guide to software pricing models.
- Perpetual licensing — one-time purchase, the customer owns the software outright
- SaaS pricing — subscription-based, the vendor hosts and maintains it, the customer pays ongoing fees
Which pricing strategy should guide your model?
Before touching a model at all, the real thinking has to go into strategy — how you land on the actual number.
- Value-based pricing — anchors the price to what the customer actually gets out of it. Takes real research into willingness to pay to get right.
- Cost-plus pricing — work out your costs, tack on a margin, done. Simple, but it tends to leave money on the table.
- Competitor-based pricing — set prices relative to the market. Fine for getting a foot in the door, risky as a long-term plan, since on its own it's a race to the bottom.
What are the most common SaaS pricing models?
There are more structures here than most people think — and increasingly, most SaaS companies combine more than one.
- Flat-rate — one price, full access, nothing metered. Simple to communicate, but it caps how much revenue can grow from each account.
- Tiered — a few packages split by feature, support level, or user limit. What most SaaS platforms are actually running under the hood, since a solo user and an enterprise buyer can both find their price point in the same product.
- Per-user — the classic charge per seat, also called seat-based pricing. Predictable, but it starts falling apart the moment value comes from AI features left open to every user as all-you-can-eat.
- Usage-based — charges strictly for what's consumed: API calls, tokens, storage. Also called pay-as-you-go or consumption-based pricing. Low barrier to entry, but revenue is less predictable.
- Freemium — free at the base tier, paid upgrades for advanced features or higher limits. Strong top-of-funnel, but conversion rates can run low.
- Per-feature — customers pay for the specific capabilities they actually turn on. Clear upgrade path, though it can feel like nickel-and-diming if the line is drawn wrong.
- Hybrid — blends a subscription with usage limits and purchasable add-ons stacked on top. Basically where everyone ends up, since it keeps the predictability of a subscription with the upside of usage growth.
How is AI reshaping SaaS pricing models?
AI features don't cost you per user — they cost you per token, per inference, per compute cycle. A flat per-seat price simply can't absorb that kind of variable cost, which is exactly why token-based and hybrid models are becoming the default for anything AI-powered rather than the exception.
None of this works, though, without a way to actually meter that usage in real time, down at the event level — you can't price what you can't measure. See more on this shift at the AI pricing engines hub.
Does SaaS pricing differ by business type?
The right model also depends on who's buying.
- B2C SaaS — usually wants a low price point, high volume, and a premium layer to drive self-served growth.
- B2B SaaS — can support higher prices and longer sales cycles, so tiered or value-based packaging tends to win.
- Enterprise SaaS — really its own category: custom contracts, annual billing, negotiated terms, and pricing that has to bend around whatever entitlements that specific customer needs.
How do you choose the right SaaS pricing model?
It comes down to one question first, then a discipline of revisiting it.
What unit really reflects the value your customer is getting — seats, usage, features, or something else entirely? Once that's clear, and you've figured out what people will actually pay for, model your costs against it — especially if you're running AI, where inference cost per customer swings around. Then plan on revisiting the whole thing at least every few months, because pricing is never really finished; some AI companies change their prices as often as every couple of weeks. For the fully detailed write-up of every model, including a side-by-side comparison table, see the complete SaaS pricing models guide.
How do you change your SaaS pricing model without a code release?
Picking a model usually isn't the hard part — changing it later is. Most companies hardcode pricing and packaging logic straight into the product, so every change to what's in which package turns into an engineering sprint instead of the two-minute configuration change it should be.
The fix is decoupling packaging and pricing configuration from the product code entirely. Licensing & Entitlements (Zentitle) enforces the tiers, seats, and feature gates in real time; Monetization Engine (Zenmeter) handles the metering and rating behind usage-based and hybrid plans. Together, under the Nalpeiron Growth Platform, that's what lets a pricing model change without an engineer touching a single line of code.
Book a demo to see your own pricing model running on the platform.
Video transcript
Auto-generated from the video and lightly edited for readability.
Most SaaS companies charge per seat, but things are changing. AI services don't charge by traditional seats because AI costs you more the more you use it. Per-seat pricing is increasingly starting to break down in the world of SaaS, driven by AI functionality and a need to align pricing models with the value delivered to the customer.
Let's start simple. A SaaS pricing model is just the structure behind how you charge. This could be by flat fee, per seat, by usage, whatever shape it takes. That's a different thing from pricing strategy, which is deciding how you land on the actual number — what value you're charging for, and why. You need both: the model structure and the strategy.
Let's see if we can unpack this a bit. Before you even touch a model, deep thinking needs to go into your strategy.
Pure value-based pricing anchors the price to what the customer actually gets out of it. It takes real research to get right.
Cost-plus is the lazy cousin. Work out your costs, tack on a margin for the costs you're incurring in infrastructure and AI tokens — done. Easy, but you're probably leaving money on the table.
Competitor-based pricing is an option — copying their logic. It's fine for getting a foot in the door, but risky as a long-term plan, because eventually it's just a race to the bottom.
Whichever way you approach it now, you might then need to pivot and change later. So you'll need the flexibility to do this. SaaS companies need to test, launch, and iterate on pricing as fast as the market moves — and the SaaS market is changing faster than ever before, driven by AI.
Okay, now the models themselves — there are more of these than people think.
Flat-rate is one price, full access, nothing metered. Simple, but it stops you growing revenue from each account.
Per-user is the classic charge per seat. Predictable, except it starts falling apart the moment your value comes from AI features left open to users as all-you-can-eat.
Tiered is what most SaaS companies are actually running under the hood — a few packages split by feature or limit, so a solo user and an enterprise buyer can both find their price point in the same product.
Usage-based flips the whole thing around and just charges for what's actually consumed: API calls, tokens, storage.
Per-feature charges based on which capabilities someone actually turns on.
And hybrid, which at this point is basically where everyone ends up, blends a subscription with usage limits and purchasable add-ons stacked on top.
And here's why all of this matters right now, more than it used to. AI features don't cost you per user — they cost you per token, per inference, per compute cycle. A flat per-seat price simply can't absorb that kind of variable cost, which is exactly why token-based and hybrid models are becoming the default for anything AI-powered. None of this can work, though, without a way to actually meter that usage in real time, down at the event level.
The right model also just depends on who's buying. B2C SaaS usually wants a low price point, high volume, and a premium layer to drive self-served growth. B2B can support higher prices and longer sales cycles, so tiered or value-based packaging tends to win — and enterprise is really its own category: custom contracts, annual billing, negotiated terms, pricing that has to bend around whatever entitlements that specific customer actually needs.
So how do you actually choose? It comes down to one question: what unit really reflects the value your customer's getting — seats, usage, features, or something else entirely? Once you know that, and have figured out what people will actually pay for, you can model your costs against it — especially if you're running AI, where inference costs swing around.
You'll need to plan on revisiting the whole thing at least every few months, though, because pricing is never really finished. Some AI companies are changing their own prices as often as every couple of weeks. When you're incurring increasing variable costs from the AI powering the AI-based features you've added into your own SaaS product, you might need to respond by adjusting your own pricing.
Here's the part that trips people up, though: picking the model isn't usually the hard part — it's changing it later that is. Most companies hardcode their pricing and packaging logic straight into the product, so every change to what's in what package turns into an engineering sprint instead of the two-minute configuration change that it should be.
The fix is to decouple packaging and pricing configuration from your product code entirely. This is exactly the problem the Nalpeiron Growth Platform solves. You can run whatever pricing model you want and actually change it without an engineer touching a single line of code — all the customer entitlements and usage metering are taken care of.
Want more detail? Visit our website and book a demo. We can show you how our platform helps you deliver SaaS pricing and packaging without those slow, expensive engineering sprints.
Decouple your
monetization today
Join the enterprises scaling their revenue without rebuilding their stack every year.
Unleash the monetization potential of your software / SaaS / Hardware