Quick answer: what changes on October 1, 2026
Meta announced that, starting October 1, 2026, it will charge for each service message delivered through WhatsApp Business Platform after a monthly free allowance. Each business phone number will receive 1,000 service messages at no charge; Meta will begin charging with the 1,001st delivered message. The rule is published in the official WhatsApp Business Platform pricing documentation.
The allowance is 1,000 service messages per business phone number each month
Charges begin with the 1,001st delivered service message
Unused free messages do not roll over to the next month
The 24-hour customer service window continues to work as before
Utility templates sent inside that window will also become chargeable
Messages customers send to businesses remain free
The threshold is not 1,000 conversations, 1,000 customers, or 1,000 messages sent by users. It is 1,000 service messages delivered by the business from each phone number during the month.
What exactly is a service message?
A service message is a non-template reply a business sends inside the 24-hour customer service window opened by the user's latest message. It can be written by a person, an automation, or an AI Agent; who creates the reply does not change its pricing category.
The window opens and resets whenever the user sends a message. Inside it, the business can deliver regular replies such as text, images, and other supported message types. Outside the window, the business can only contact the user again with an approved template that matches its purpose.
Incoming customer message: no Meta charge
Regular business reply inside 24 hours: service message
Utility template: keeps its own category even when sent inside 24 hours
Marketing or authentication template: keeps its applicable category and rate
How the 1,000-message allowance works
The allowance is calculated per business phone number. If a business has several numbers, each receives its own monthly allowance of 1,000 service messages. Unused messages do not carry over, and the count resets each month.
Meta charges on delivery, not on the send attempt. Cost control should therefore measure how many replies were actually delivered and which pricing category was recorded, not only how many responses the system tried to produce.
Up to 1,000 service messages delivered from one number during the month: no charge
Starting with message 1,001: each delivered service message is chargeable
The allowance does not cover utility, marketing, or authentication templates
Adding human agents, automations, or AI Agents does not multiply the allowance
Adding phone lines only to chase allowances creates complexity and does not replace sound operational architecture
Which categories are charged after the change
Service: 1,000 free messages per number each month; additional messages are charged per delivery
Utility: every delivered template is charged, including inside the 24-hour window; it does not use the service allowance
Marketing: continues to be charged for every delivered template
Authentication: continues to be charged for every delivered template and may have an international rate when applicable
Incoming messages: messages users send to the business are not charged
The rate depends on the recipient WhatsApp phone number's country calling code and the category. Utility and authentication pricing may also use Meta's published volume tiers.
The 72-hour free-entry window still matters
When a person starts a conversation from an ad that clicks to WhatsApp or a Facebook Page call-to-action button, and the business replies within the first 24 hours, a 72-hour free-entry-point window opens. During that window, Meta does not charge for supported messages, including templates.
This free window is independent of the customer service window. If the 24-hour window closes while the 72-hour window remains open, the business can send only templates, even though those templates may remain free during the free-entry window. The strategy should respect customer intent; a pricing exception is not a reason to send unnecessary messages.
How much does it cost after message 1,000?
Meta published that the service rate will equal the list rate for utility and authentication messages in each market. There is no single worldwide rate: the amount depends on the recipient's country or region and the billing currency.
The monthly formula for each phone number is: billable service messages = the greater of delivered service messages minus 1,000 or zero. Multiply the billable messages by the applicable service rate for the recipient's market. You can run the calculation with OneChat's WhatsApp Business API pricing calculator.
Always use the current official rate card. Rates can change and are based on the recipient's market, not simply the country where the business is registered.
Colombia example
For messages delivered to Colombian numbers, Meta's published October 1, 2026 rate card lists a service rate of COP 2.9455 per message after the allowance. The following examples assume one business phone number and recipients with the +57 country code:
900 service messages delivered in the month: 900 free and 0 billable
1,250 delivered messages: 1,000 free and 250 billable; estimated Meta charge: COP 736.38
5,500 delivered messages: 1,000 free and 4,500 billable; estimated Meta charge: COP 13,254.75
These examples explain the mechanism and do not replace Meta's invoice. If the operation serves recipients in several markets, each billable message must use the rate that applies to the recipient's number.
Best practice 1: measure delivered messages, not only conversations
One conversation may contain one reply or twenty. Under the new model, two operations serving the same number of customers can have different costs if one fragments every answer into many bubbles while the other replies more completely.
Track delivered service messages by number and month
Separate send attempts, deliveries, and failures
Store the pricing category reported by Meta
Segment consumption by recipient market
Compare traffic handled by people, automations, and AI Agents without confusing messaging charges with AI model usage
Best practice 2: answer completely without making the conversation robotic
Avoid automatically splitting one response into a greeting, confirmation, explanation, and closing as four messages when it can be communicated clearly in one or two. Each delivered fragment can consume the allowance or create a charge.
Optimization does not mean sending endless paragraphs. A good reply remains concise, scannable, and useful. Remove artificial fragmentation, filler messages, and repeated confirmations without sacrificing comprehension.
Group information that belongs to the same idea
Do not send a separate message only to say that the system is searching
Use short lists when they improve readability
Request the information genuinely needed in one intervention
End with one clear next step instead of several open questions
Best practice 3: coordinate bots, AI, and the human team
Duplicate responses cost money and damage the experience. If an AI Agent replies while a person intervenes, or if an automation retries without control, the business may pay for messages that provide no value.
Pause automation when a human agent takes the conversation
Prevent two flows from responding to the same event
Use idempotency so a retry does not duplicate an already delivered reply
Define ownership for every stage
Hand off to the human team with collected context so customers are not asked for the same information again
Best practice 4: design to resolve inside the 24-hour window
The pricing update does not change the customer service window. Resolving the customer's intent while it remains open is still useful: identify the need, consult authorized information, confirm the result, and transfer with context when necessary.
Prioritize first-contact resolution
Do not send artificial replies to try to keep the window open
When later outreach is necessary, use an approved template in the correct category
Schedule reminders only when customers expect them and they support a useful next step
Restore context without forcing the customer to repeat the entire request
Best practice 5: classify templates correctly
Starting October 1, a utility template sent inside the 24-hour window will no longer be free. It should not be presented as a service message or broken into regular replies to evade its category. Meta may review and recategorize templates according to their content and purpose.
Utility: a specific, non-promotional update related to a user action or request
Marketing: promotions, offers, commercial recommendations, or re-engagement
Authentication: supported one-time passwords and verification use cases
Service: a non-template reply inside an open customer service window
Before launching campaigns or automations, review the WhatsApp Business Messaging Policy, preserve consent when required, and honor opt-out requests.
Best practice 6: create alerts before message 1,001
A business should not discover the charge at the end of the month. Configure a view for each phone number that shows allowance consumption, the projected month-end volume, and the cost of additional messages.
Early alert near 70% of the allowance
Operational review at 90%
Projection of billable messages at 100%
Separate budgets for service, utility, marketing, and authentication
Compare cost with resolution, satisfaction, and sales so volume is not the only optimization target
Best practice 7: audit delivery statuses and billing
Meta reports in status webhooks whether a message is billable and which pricing category was applied. Store that data with the message identifier, business number, market, and date. This traceability helps explain differences between internal volume and the invoice.
Do not count a failed attempt as a charged delivery
Reconcile delivery events with Meta's reports
Investigate retries, duplicate responses, and unexpected category changes
Keep Meta messaging charges, the OneChat plan, and AI provider usage separate
Review the configuration after changing a flow, template, or phone number
September 2026 preparation checklist
Inventory every business phone number and its owner
Measure at least the last 60 to 90 days of delivered service messages
Identify which numbers exceed 1,000 messages each month
Forecast costs by market and category using the October rates
Review whether the bot or AI fragments replies
Validate that automation pauses during human service
Audit utility, marketing, and authentication templates
Store the category and billable status reported by Meta
Define alerts and a monthly budget
Test scenarios in OneChat's calculator before the effective date
The right preparation is not to stop responding. It is to know what is delivered, why it is sent, what it costs, and whether each message genuinely helps resolve the customer's intent.
Frequently asked questions
Are the first 1,000 conversations free? No. The allowance covers 1,000 service messages delivered from each business phone number during the month.
Can a reply written by a person be charged? Yes. After the allowance is exhausted, a delivered service reply can generate a charge even when a human agent writes it.
Are customer messages charged? No. Messages users send to the business remain free.
Does the allowance cover utility templates? No. The allowance applies to service messages. Utility templates sent inside 24 hours will also be charged starting October 1, 2026.
Do unused messages roll over? No. The allowance does not accumulate and resets monthly for every phone number.
Did the 24-hour window change? No. A service message can only be sent while the customer service window is open.
Is the rate the same in every country? No. Meta sets market and category rates using the recipient's country calling code.
Official sources and tools
Official WhatsApp Business Platform pricing documentation, updated by Meta on August 5, 2026
Rate cards effective October 1, 2026
WhatsApp Business Messaging Policy
OneChat WhatsApp Business API pricing calculator
Review Meta's pricing documentation, the official messaging policy, and OneChat's calculator. Rates and rules can change; confirm the current version before making budget or technical decisions.
