diff --git a/crates/formula/src/lexer.rs b/crates/formula/src/lexer.rs
index 83ea78a86..332cb6eeb 100644
--- a/crates/formula/src/lexer.rs
+++ b/crates/formula/src/lexer.rs
@@ -346,7 +346,7 @@ pub fn tokenize_locale(src: &str, locale: crate::FormulaLocale) -> Result<Vec<To
                     // `Sheet1!A1:Sheet1!B2` is a range whose second end repeats the sheet, not a span.
                     let repeats_sheet = is_range_end(&w)
                         && matches!(out.last(), Some(Token { tok: Tok::Sheet(p, None), .. })
-                            if chars.get(j + 1..k).is_some_and(|s| s.iter().map(|c| c.1).eq(p.chars())));
+                            if chars.get(j + 1..k).is_some_and(|s| crate::same_name(&s.iter().map(|c| c.1).collect::<String>(), p)));
                     k > j + 1 && ch(k) == Some('!') && !repeats_sheet
                 } {
                     let mut k = j + 1;
diff --git a/crates/formula/src/tests.rs b/crates/formula/src/tests.rs
index 2318a8b29..a7506aa4b 100644
--- a/crates/formula/src/tests.rs
+++ b/crates/formula/src/tests.rs
@@ -10,6 +10,8 @@ fn range_end_repeating_sheet_name_round_trips() {
     assert_eq!(rt("'a'!A1:'a'!B2"), "a!A1:a!B2");
     assert_eq!(rt("Sheet1!A1:Sheet1!B2"), "Sheet1!A1:Sheet1!B2");
     assert_eq!(rt("SUM(a!A1:a!B2)"), "SUM(a!A1:a!B2)");
+    // Sheet names match without regard to case, as Excel's do.
+    assert_eq!(rt("Sheet1!A1:SHEET1!B2"), "Sheet1!A1:SHEET1!B2");
 }
 
 #[test]