Throughput in SMS is normally quoted in messages per second, and the question "what is my throughput" has two answers that people routinely mix up. One is about your relationship with the API. The other is about the network's relationship with the handset.
What are the two limits?
An API rate limit you hit while submitting, and a carrier ceiling that governs how fast accepted messages actually leave.
The API rate limit governs how many requests you may make. Bird's rate limits guide covers how a limit is keyed, how it is resolved for a given caller, and what the response headers report. Exceeding it is loud and immediate: you get a 429 and nothing was sent. What is API rate limiting and how do I handle a 429 covers the handling.
The carrier throughput ceiling governs how fast messages move from the platform toward handsets in a given destination. It is a property of the route rather than of your account, and exceeding it does not produce an error. The messages are accepted and simply take longer to leave.
That difference in failure mode is the useful part. If you are getting 429s you are calling too fast. If your API calls all succeed and delivery lags, you are past a network ceiling and no amount of retrying will help.
What sets the carrier ceiling?
The destination and the sender type together, which is why there is no single number.
Bird publishes throughput per destination as data rather than in prose: the country policy carries a maximum throughput per second, and each sender type available in that country carries its own messages-per-second figure. The two are separate because a short code and a long code in the same country do not move traffic at the same rate.
The current figures for any destination are on the SMS destinations page, which reads them live. They are not reproduced here, because a throughput number copied into prose is wrong the moment a carrier changes it and nothing tells you.
The shape to take away is that throughput is attached to where you are sending and what you are sending from, not to your account or your plan.
Why is my campaign slower than my sender's rated throughput?
Because the rated figure is a ceiling on a route, not a promise about your traffic.
Several things sit between the two. Messages to different destinations move at different rates, so a mixed international campaign runs at a blend rather than at its fastest route's speed. Messages that split into several segments occupy the route several times over, so a two-segment body halves your effective message rate. And the message centre is store-and-forward, so a handset that is unreachable holds its message without holding up others, but does mean "sent" and "delivered" drift apart. What is an SMSC covers that machinery.
The practical reading: throughput tells you how fast messages can leave, and says nothing about when they arrive.
How do I get more?
Usually by changing the sender, not by changing your code.
Since the ceiling is attached to the sender type in a destination, the lever is which sender you use. A short code exists largely because it carries volume that a long code cannot, and choosing between sender types is a throughput decision as much as a cost or trust one. Which SMS sender type should I use covers the trade in full, and per-country availability is on the destinations page.
Two things that do not help, and are commonly tried. Sending faster does not raise a network ceiling; it fills a queue. And spreading the same campaign across many numbers to multiply throughput is snowshoeing, which carriers treat as an abuse pattern rather than as scaling, and which the CTIA messaging principles name explicitly.
If you are planning a send large enough for throughput to matter, the useful preparation is arithmetic rather than tuning: total segments divided by the route's messages per second gives the floor on how long the send takes, and what is bulk SMS covers the rest of the planning.