SMS character limit: how many characters fit in one text

The 160 figure comes from ETSI TS 123 038, the standard that defines how SMS text is coded. For the GSM 7-bit default alphabet, the standard says "the message can consist of up to 160 characters."

The standard gets to 160 by counting bits. A message can use up to 140 octets, and an octet is 8 bits, so the message has 1,120 bits to work with. Each GSM 7-bit character takes 7 of them, so (140 × 8) ÷ 7 = 160 characters. The same 140 octets hold only 70 UCS-2 characters, because each one needs more bits. Clause 6.1.2.1.1 of the standard makes this calculation.

The character set decides which of those numbers applies. Basic letters, digits and common punctuation are in the GSM 7-bit alphabet. Emoji and many accented or non-Latin letters aren't, so one of them pushes a message into UCS-2. Twilio's UCS-2 glossary says a message falls back to UCS-2 if it includes any character outside GSM-7. A single emoji is enough to move the whole message from 160 to 70, even if the rest of the text is plain.

Some characters come from the extension table of the GSM 7-bit alphabet, and each one takes two places. The standard says the displayable count drops by one for every use of that table. Twilio's GSM-7 glossary says these escape characters count as two characters. Take a message with 155 visible characters, 6 of them from the extension table. The other 149 take one place each and the 6 take two, so the message needs 161 places, one over the 160 limit. The counter reads 155 and still shows room.

A message longer than the single-message limit is split into parts, and each part carries a header. Twilio says the common 6-byte header leaves 153 GSM-7 characters or 67 UCS-2 characters for the text in each part. Carriers and message types can differ, so treat those as the common numbers, not a promise.

The counter tracks the 160 limit, and that limit applies only to text in the GSM 7-bit alphabet. It counts visible characters, so an emoji counts as one. It can't tell which character set your text needs, and it doesn't weigh extension characters, count parts or detect a switch to UCS-2. For how emoji are counted in general, see emoji and Unicode counts.