Unicode. Internally Squeak uses UTF-32. Use ASCII only – your strings will be instances of ByteString and contain characters of one byte each. Use any character with Unicode code point of 256 or above – your strings will be instances of WideString and contain characters of 4 bytes each.
Squeak contains conversion methods for converting UTF-8 to its internal representation and back so UTF-8 can be used when importing/exporting files, talking to external libraries and the clipboard.