SMS character counter

The short answer

An SMS carries 140 bytes of text. In the GSM-7 alphabet each character uses 7 bits, so 160 characters fit. In Unicode (UCS-2) each character uses 16 bits, so 70 fit.

A longer message is split into parts that the phone joins again. Each part gives 6 bytes to a joining header, so it carries 153 GSM-7 characters or 67 Unicode characters.

Encoding	One SMS	Each part of a long SMS	Used for
GSM-7	160 characters	153 characters	English letters, digits, common symbols and some accented letters
Unicode (UCS-2)	70 characters	67 characters	Arabic, Cyrillic, Hindi, Chinese, emoji and any other character outside GSM-7

Sources: ETSI TS 123 038 (3GPP TS 23.038), section 6.1.2.1.1 ("in 140 octets, it is possible to pack (140x8)/7=160 characters") and 6.2.3 (UCS2, 16 bits per character); 3GPP TS 23.040, 9.2.3.24.1 (concatenated short messages). Checked 6 November 2026.

How the counting works

GSM-7 first. The counter checks every character against the GSM 7-bit default alphabet in 3GPP TS 23.038. It has 128 positions: English letters and digits, common punctuation, some accented and Greek letters such as é, ñ, Ü and Δ, and an escape code.

Extension characters count as two. Nine more symbols, plus a page-break control, live in the GSM-7 extension table: ^ { } \ [ ~ ] | €. Each is sent as an escape code plus the character, so each uses two places.

One outside character switches everything. If any character is outside both tables, the whole message goes in Unicode (UCS-2). Then every character counts, including spaces and digits, in 16-bit units.

Long messages. Above 160 (GSM-7) or 70 (Unicode), the counter fills parts of 153 or 67 in order. It never splits a two-unit character, such as € or an emoji, across two parts, so a part can end one short.

Why 153 and 67? The joining header is 6 bytes: its length, an identifier, the data length, a reference number, the number of parts and the part number. In GSM-7, 6 bytes (48 bits) plus one fill bit take the space of 7 characters, leaving 153. In Unicode, 134 bytes are left, which is 67 characters.

Sources: 3GPP TS 23.038, 6.2.1.1 (extension table); 3GPP TS 23.040, 9.2.3.24 (user data header and fill bits); Unecast SMS API reference (6-byte header, 153 and 67 per part). Checked 6 November 2026.

Other languages, accents and emoji

• Some accented letters fit GSM-7. é, è, ù, ì, ò, à, ñ, Ñ, ü, ä, ö, ß, Å, Æ and Ø are in the GSM-7 alphabet, so many French, German, Italian and Nordic messages still fit 160 characters.
• Others force Unicode. á, í, ó, ú, ã, õ, â, ê and Turkish ı, ş and ğ are not in GSM-7, so one of them switches a Spanish, Portuguese or Turkish message to Unicode. Try the Español sample above.
• Greek capitals such as Δ, Ω and Σ are in GSM-7. Lowercase Greek is not.
• Non-Latin scripts such as Arabic, Cyrillic, Hindi, Chinese and Thai are always sent in Unicode: 70 characters in one SMS, 67 per part.
• Vowel signs count separately. In scripts such as Hindi, a letter with a vowel sign is two Unicode characters. For example, हि is ह plus ि, so it counts as two.
• Mixing languages does not help. One Arabic or Hindi word in an English message turns the whole message into Unicode.
• Emoji count as two units each, and emoji with skin tones or other modifiers count more.
• National language tables. 3GPP TS 23.038 also defines national language shift tables, for example for Turkish, Spanish, Portuguese and several languages of India, which keep some of these characters in 7 bits. Support depends on the network and the phone, so this counter uses the default GSM-7 alphabet and its extension table.

How to stay in one part

• Watch for hidden Unicode. Curly quotes (’ “ ”), long dashes (– —), the ellipsis (…) and no-break spaces often arrive when you paste from a word processor. Use the replace button above to swap them for plain characters.
• Count € and brackets twice. Each of ^ { } \ [ ~ ] | € uses two places in GSM-7.
• Keep emoji for marketing. One emoji cuts a single SMS from 160 to 70 characters.
• Write the short version first. Put the key fact, such as the code or the amount, at the start.
• Split by language. Send a non-Latin script and English as separate messages if each then fits in fewer parts.

Parts and cost

Each part of a long message travels as its own short message, and the phone joins them. Check how your provider bills multi-part messages.

On Unecast, the SMS portal's message details show each message's part count and cost. Current SMS rates are on the bulk SMS gateway page. If you send over SMPP, set data_coding to 0 for GSM-7 or 8 for Unicode, as the SMS API reference shows.

Frequently asked questions