How many characters fit in one SMS
The 3GPP standard for text messages, TS 23.038, defines the GSM 7-bit default alphabet: Latin letters, digits, common punctuation and a few accented letters such as é, ñ and ü. A message written only in that alphabet can hold up to 160 characters. A few symbols, namely € [ ] { } | ^ ~ and the backslash, sit in an extension table and are sent as two characters each. When a message is longer than one SMS, the phone sends it in parts that carry a small header so they can be joined again, and the companion standard TS 23.040 sets the room left in each part at 153 characters.
Why Arabic, Urdu and emoji cut an SMS to 70 characters
TS 23.038 states that Arabic text cannot be written in the GSM 7-bit alphabet, so the phone switches the whole message to UCS-2, which uses 16 bits per character. The same 140 bytes then hold only 70 characters, and 67 per part when the message is split. One Arabic letter or one emoji is enough to switch an English message too, and so is a curly apostrophe or an accented letter pasted from a document. The counter lists those characters so you can replace them. Most emoji need two 16-bit units, so each one uses two places. Some networks use the national language tables the standard also defines, including one for Urdu, and bulk SMS services may count slightly differently, so check with your provider before a large send.
How X counts 280 characters
X does not simply count letters. Its open-source counting rules give most Latin, Arabic, Cyrillic and other alphabetic characters a weight of 1, and Chinese, Japanese and Korean characters and most other symbols a weight of 2. Every emoji counts as 2 however many code points it is built from, and every link counts as 23 whatever its real length. The counter applies those rules after normalising the text the same way. It recognises links that start with http or www; a bare domain written without them may also be counted as a link by X.
Meta titles, descriptions and video fields
Google says there is no limit on the length of a page title or a meta description; it cuts them in search results as needed to fit the width of the device. The counter therefore shows their length without a pass or fail mark, and figures such as 60 or 155 characters are rules of thumb, not Google limits. YouTube does set hard limits: 100 characters for a video title and 5,000 for its description, according to YouTube Help.
Characters, code points and bytes
Characters here are what a reader sees, so an emoji with a skin tone or an Arabic letter with its vowel mark counts once. Code points are what many databases and programming languages count, and they are often higher. UTF-8 bytes are what a file or a database field actually stores: an English letter takes 1 byte, an Arabic letter 2, the euro sign 3 and most emoji 4. Byte limits matter for database columns, file names and some forms. To count words, sentences and reading time instead, use the word counter.