WhatsApp will charge for every message from October 1. The bill lands hardest on chatty automation.

From October 1, 2026, Meta charges for every message a business sends on the WhatsApp Business Platform. Its pricing page: "Meta will charge on a per-message basis for all service messages, consistent with how Meta charges for template messages. These messages have not been charged since November 1, 2024." [1] Service messages are the replies you write in your own words inside the 24-hour customer service window. That is where support teams and automated replies do most of their work. Utility templates sent inside that window, free since July 2025, are billed too. [1] Meta has since published the October 1 rate card with service rates, and the terms of a free monthly tier of 1,000 delivered service messages per phone number. [7]

What moves your bill now is how many messages your business sends per conversation.

We run AI agents on WhatsApp for businesses across Southeast Asia, so this lands on us and on our customers. This post checks the change against Meta's own docs and prices it at real conversation volumes.

Updated 21 September 2026: the Southeast Asia rates now sit in two tables, in dollars and in ringgit, with each market's marketing rate beside its service rate.

Updated 16 September 2026: Meta has published the free tier's terms, and its two pricing pages now disagree about what happens to an account with no payment method. [1][7] The free tier section, the deadline, the checklist and the FAQ below reflect both, and the figures now carry Meta's ringgit rates next to the dollar ones. An earlier update on 3 September added the October rate card with service rates.

What changes on October 1

Every message a business delivers becomes billable. Message type by message type:

What you send	Today	From October 1, 2026
Free-form reply inside the 24-hour window (a service message)	Free	Billed per message [1]
Utility template inside the 24-hour window	Free	Billed per message [1]
Utility template outside the window	Billed per message	Billed per message, with rate card updates in some markets [2]
Marketing template	Billed per message	Billed per message
Authentication template	Billed per message	Billed per message
Reply powered by Meta Business Agent	Billed per token since August 1, 2026	Billed per token [1]
Anything a customer sends you	Free	Free

Most coverage skips the second half of the change, which is about utility templates. Effective October 1, 2026, Meta "will charge on a per-message basis for utility messages sent in response to users (within an open 24-hour customer service window). These messages have not been charged since July 1, 2025." [1]

Service messages are billed per message, at the same rate as utility and authentication templates in each country. [1] They also miss the volume discount that templates get: "Volume tiers: None. Meta will not offer volume tiers for service messages. Meta continues to offer volume tiers for utility and authentication messages." [1] You pay the same rate on your first reply of the month as on your hundred-thousandth.

One deadline sits before October 1

Everything else in this change costs money. This one stops your replies going out: "For any Solution Provider or directly-integrated businesses that does not have a payment method on file by September 30, 2026, Meta will stop delivering service messages as of when they become charged on October 1, 2026." [1]

Meta's main pricing page, updated on September 10 and so newer, sets a softer rule: "If you do not have a payment method for your WhatsApp Business account, Meta will deliver service messages within the shared free tier but not deliver them after the free tier has been used." [7] The two pages describe the same month differently, and we plan against the stricter one. Either way, a business that never adds a payment method does not get a bigger bill; it goes quiet, on October 1 under the older page, or once the number has used its 1,000 free service messages under the newer one. If you are directly integrated, add the payment method this month. If a provider sits between you and Meta, ask them today whose payment method is on file.

Two kinds of non-template message, two different meters

Since July 1, 2026, a non-template reply falls into one of two categories, metered differently. [1] The split matters most when you are choosing an AI vendor.

Category	Who powers it	How Meta charges
Service	A human agent, or a third-party AI solution	Per message, at the market's utility rate, from October 1, 2026 [1]
Meta Business Agent	Meta's own agent platform	Per token, since August 1, 2026, at 2 US dollars per million tokens [1]

Meta's definition: "Any non-template message that is not powered by Meta Business Agent is a service message," and service messages "can be powered by a person, like a customer service representative, or by a 3rd-party AI solution." [1] An AI that runs outside Meta produces service messages at the flat per-message rate, and you pay your AI provider separately. Meta's own agent bills the AI and the delivery together, per token, which Meta puts at roughly 4 to 5 US cents per message at 20,000 to 25,000 tokens. [1]

Meta publishes a comparison for 1,000 AI-powered replies in Brazil: a lower-complexity third-party setup lands around 27 dollars, a higher-complexity one around 97 dollars, and Meta Business Agent around 40 to 50 dollars. [1] Which one wins depends on your model and the questions you answer. Both sides publish a meter now, so you can run the numbers for your own traffic.

The 24-hour window still decides what you can send

We keep seeing "the 24-hour window is dead." It is not. The window has always done two jobs, and it loses one of them.

When a customer messages you, a 24-hour window opens. While it is open, you can reply in your own words. Once it closes, you can only send an approved template. That is the first job, and it does not change. Meta's docs still say a non-template message can only go out while the window is open. [2]

The second job was pricing. A reply sent inside the window was free, and anything sent outside it was paid. That is what ends on October 1. Your reply now costs the same whether the window is open or closed.

So the work that goes into opening windows, keeping them warm and sending before they close still decides what you are allowed to send. It no longer saves you money. Sending fewer messages does.

How the pricing got here, in three steps

October 1 is the third time Meta has put a price on something that used to be free.

Date	What changed
November 2024	Meta stops charging for service conversations. The reply lane becomes free [1]
July 1, 2025	Conversation-based pricing ends. Templates bill per message, and utility templates inside an open window become free [1][2][4]
August 1, 2026	Meta Business Agent, Meta's own AI for businesses, moves to usage billing at 2 US dollars per million tokens [1]
October 1, 2026	Service messages and in-window utility templates become billable, with a free monthly tier of 1,000 service messages per phone number. Every business-sent message beyond that carries a price [1][7]

First conversations went per-message, then Meta's own AI went per-token, and now the reply itself goes per-message. We plan on the number of messages you send staying a permanent cost of using WhatsApp.

What it costs, in numbers

Rates are per message and set by the recipient's country, not by where your business sits. On the October 1 card, service messages cost exactly the utility rate in every market, so one column covers all three non-marketing categories. [7]

Southeast Asia, in US dollars:

Recipient market	Service, utility, authentication	Marketing
Malaysia	$0.0140	$0.0860
Singapore	$0.0160	$0.0732
Indonesia	$0.0250	$0.0411
Philippines, Thailand, Vietnam	$0.0113	$0.0842

The Philippines, Thailand and Vietnam have no row of their own. They sit in Meta's "Rest of Asia Pacific" band, below every individually priced market in the region, so a Bangkok or Manila business is better off for the omission. [2][7] For that band, October raises marketing from 0.0732 dollars. [7]

Meta bills each account in its own currency, so an account billed in ringgit reads the MYR card rather than converting the dollar rates. [7]

Recipient market	Service, utility, authentication	Marketing
Malaysia	RM 0.0564	RM 0.3467
Singapore	RM 0.0646	RM 0.2951
Indonesia	RM 0.1008	RM 0.1657
Philippines, Thailand, Vietnam	RM 0.0456	RM 0.3394

Authentication-international is priced apart from the three columns above, where it applies at all: 0.0418 dollars for Malaysia and 0.1360 for Indonesia, with no rate on the card for Singapore or the band. [7] Meta publishes both cards as downloads from its pricing page, so you can check any vendor's table against them, ours included. [3][7]

The free tier's terms are published too. "Each business phone number has one shared free tier of 1,000 delivered service messages per month," starting on October 1. [7] Unused messages do not roll over, and the tier resets monthly, 1,000 for October and another 1,000 for November. Group chats draw on the same allowance: "a group send consumes one unit per delivered recipient." [7] The tier covers service messages only, so a utility template sent inside that same open window is billed from its first message. [7] The allowance sits with the phone number, and it only ever covers the replies you write inside an open window.

You can count the tier down as the month runs rather than waiting for the invoice. From October 1, Meta's status webhooks mark a service message inside the tier as not billable, with the pricing type free_customer_service, and a message sent after the tier is used as billable, with the type regular. [7] Anything that stores those webhooks can meter the free tier live.

You cannot change what a message costs. You can change how many messages it takes to finish a conversation, and that number varies a lot.

Conversation pattern	Business-sent messages	Cost in Malaysia	Cost in Indonesia
A support question answered in one grounded reply plus a confirmation	2	$0.028	$0.05
The same question through a menu bot: greeting, menu, re-prompt, answer, satisfaction ask	6	$0.084	$0.15
A lead qualified across a day of back-and-forth, 8 business turns	8	$0.112	$0.20

At a thousand support conversations a month, the first row costs 28 dollars in Malaysia and the second costs 84.

The free tier widens that gap. A thousand dense conversations is 2,000 service messages on one number, the first 1,000 free, so the bill is 14 dollars in Malaysia. The menu-bot version sends 6,000 messages, 5,000 of them billable, or 70 dollars. Under roughly 500 conversations a month on a number, a business that answers in two messages pays nothing, while the same volume through a menu bot still pays.

None of this kills the channel. A support thread that costs a few US cents is cheap service, well under a voice call. It puts a price on sloppy design. A chatty conversation uses three to four times the messages of a tight one, so it costs three to four times as much, every month. Under the free tier it can be the difference between paying nothing and paying.

What stays free

Three things survive the change.

First, conversations that start from a Click-to-WhatsApp ad or a Facebook Page call-to-action button keep their free 72-hour entry point window, which Meta confirms is "unchanged for message delivery." [1] Four conditions are easy to miss. The customer has to arrive from the Android or iOS app, because desktop and web do not qualify. The 72 hours do not start at the click; they start when you reply, and only if you reply inside the first 24 hours. Free does not mean you can send anything. The 24-hour window runs separately from the 72, so when it closes you are back to templates only, and those stay free until the 72 hours are up. [8] And the exemption covers delivery, not Meta's own AI, whose token charges still apply. [1][8] Within those limits it is free, and you enter it by buying ads.

Second, the free WhatsApp Business app, the one a shop owner runs by hand on a phone, is untouched. [5][6] The charge applies to the API, which is what businesses use to send at scale and to automate replies.

Third, the free tier is drawn per business phone number rather than per business. [7] A shop with one number and a few hundred conversations a month sits inside it. A platform pushing tens of thousands of automated replies through one number burns it in a day.

Put together: Meta charges for automated volume, in proportion to how much you send, and leaves conversations that came from an ad alone.

Per-message pricing is a tax on chattiness

WhatsApp used to bill by the conversation. From October 1 it bills by the message, so anything that adds messages without adding outcomes adds cost.

The usual sources:

• Menu trees. A greeting, a numbered menu, a "press 2 for billing," a re-prompt when the customer types a sentence instead of a number. Four messages before any value moves, on every conversation.
• Multi-bubble habits. Automation that splits one answer across three bubbles for rhythm. That is three messages at three times the price of one.
• Re-prompt loops. Every "sorry, I didn't get that" is a message you now pay for.

Calendar drips already cost you this, and they cost more. A nurture sequence that fires on a timer reaches a customer who has not written to you, so there is no open window and no service message involved. Those are marketing templates, the most expensive category on the card: 0.0860 dollars per message in Malaysia on the October card, RM 0.3467 in ringgit, more than six times the utility rate, and 0.0411 dollars in Indonesia. [3][7] A daily touch across a five-thousand-contact list is 150,000 marketing messages a month, roughly 12,900 dollars in Malaysia, and none of that is new in October. We argued against drips when the case was about buyer attention; the invoice was already there, and October extends the meter to the replies those campaigns provoke.

A cheaper design answers the question in one message instead of five. It qualifies the lead and books the meeting in the same exchange. Buttons turn three rounds of clarifying into one tap. Follow-ups fire when the customer does something, not when a timer does. And when an interaction runs long, it moves to a page you own, with WhatsApp as the doorbell.

From October, the difference between that design and a chatty one shows up on an invoice.

What to do before October 1

Measuring takes an afternoon. Some of the changes below take longer. They are ordered by how much they cut the bill, except the first, which is here because skipping it stops your replies going out.

1. Confirm a payment method is on file before September 30. Meta's non-template pricing page says that without one, your service messages stop on October 1 rather than being billed to you. [1] Its main pricing page, updated later, says they keep flowing until the number has used its 1,000 free service messages for the month, then stop. [7] The two pages disagree, and adding the payment method settles it either way. If a provider sits between you and Meta, get them to tell you in writing whose payment method covers your account.
2. Measure messages per conversation. Pull last month's numbers and compute business-sent messages divided by conversations, per flow. This one ratio is your exposure. Anything above 4 or 5 on a support flow deserves a look at where the turns go.
3. Price your own traffic. Take last month's service-message count per number, subtract the 1,000 free, and multiply the rest by your market's rate on the October card, in the currency your account is billed in. [7] No template draws on the free tier, and a utility template sent inside the window is billed from its first message. That is roughly your October line item at today's behavior. Decide now whether it changes your product pricing, your margins, or neither.
4. Kill the free-window assumptions in your automation. Any flow that pads messages because "in-window is free" is about to convert padding into spend. Batch nudges into one daily digest instead of five singles, and merge confirmations into the answer they confirm.
5. Collapse menus into answers. If the first customer message tells you what they want, answer it. Route by understanding, not by numbered menu.
6. Move long interactions off the metered channel. Deep-link to the booking page, the order status page, the portal. One message that opens a surface you own replaces ten that reproduce it bubble by bubble.
7. Keep Click-to-WhatsApp where it fits. If you already run CTWA for lead generation, its 72-hour free window is now a pricing advantage as well as an attribution one, within the four limits above. [1][2]
8. Read the October card, meter the free tier from your status webhooks, and ask your provider how the charge reaches you. [1][7] The webhooks carry a pricing type per message, so what is left of a number's free tier is countable rather than guessable. The billing lands at the platform layer, so every BSP and every messaging vendor inherits it. What varies is how visibly it is passed through.

Common misconceptions

These are circulating in vendor posts and group chats.

"WhatsApp is now too expensive for support." A resolved support thread costs a few US cents at the published rates, and the first 1,000 delivered service messages a month on each number cost nothing. [7] What gets expensive is a design that spends six messages doing one message of work.

"The 24-hour window is gone." The window's permission rules are unchanged: free-form content still requires an open window, and templates are still the only option outside it. [1][2] What ended is the free ride inside it.

"Volume discounts will soften it." Utility and authentication templates have volume tiers. Service messages do not, and Meta says so in as many words. [1] The way to lower the service-message bill is to send fewer service messages.

"Using AI to reply means paying Meta for AI." Only if the AI is Meta's. A reply generated by a third-party system is an ordinary service message at the market rate, and Meta notes that businesses using third-party AI "typically incur separate charges for AI usage from the AI provider and message delivery from the messaging solution provider." [1]

"My provider's plan absorbs this." The charge originates with Meta and applies platform-wide, so it reaches every provider's cost base on the same day. Some will pass it through as a line item, some will fold it into plans and raise them. Either way it arrives on your invoice; ask your vendor to show you where.

Where this leaves AI agents on WhatsApp

We sell into this market, so weigh the next two paragraphs accordingly.

StaffOS sells AI staff that work over WhatsApp: support agents that resolve tickets and lead-gen agents that qualify and book. In Meta's taxonomy we are a third-party AI solution, so every reply our agents send is an ordinary service message at the market rate, and our model costs are ours to answer for. [1] Every model call we make is already billed per token and metered per tenant down to the message. From October the channel bills the same way the models do.

The pricing rewards agents that finish the job early. An agent that reads the customer's first message and answers it from your catalog does in one turn what a menu tree does in five. An agent that qualifies and books inside the same conversation does not need a drip sequence to herd the lead back four times. A follow-up that fires on a real signal instead of a schedule sends one message worth sending. If you are comparing vendors this quarter, ask each of them: what is your messages-per-resolution, and what does a conversation cost me at that cadence?

This is the second time this year the platform moved under everyone, after the identity change that took customer identity off the phone number in April. Both times the same habits paid: build against the primary contract, read the platform docs before the vendor summaries, and treat a free lane as a phase.

What to verify before you budget

The October 1 rate card is published, service rates included, and so are the free tier's terms. [7] What stays open is which of Meta's two pages is right about an account with no payment method, and adding one before September 30 takes that question off your list. [1][7] Before you commit numbers to a budget, check three things against the live docs: the October card for every country you message, in the currency your account is billed in [7], whether any volume tier applies to your utility and authentication traffic [2], and how your BSP or platform vendor surfaces the new charges on their invoice. The figures here reflect Meta's documentation and its October 1 rate cards as of 16 September 2026.

If a number here no longer matches the live rate card, tell us and we will fix it.