diff --git a/grammar/abnf-grammar.txt b/grammar/abnf-grammar.txt
index d6d4b6902a..668e8134dc 100644
--- a/grammar/abnf-grammar.txt
+++ b/grammar/abnf-grammar.txt
@@ -645,7 +645,7 @@ symbol = "!" / "&&" / "||"
        / "[" / "]"
        / "{" / "}"
        / "," / "." / ".." / "..." / ";" / ":" / "::" / "?"
-       / "->"
+       / "->" / "_"
 
 ; Everything defined above, other than comments and whitespace,
 ; is a token, as defined by the following rule.
diff --git a/parser/src/parser/context.rs b/parser/src/parser/context.rs
index 1b995b7335..a43bb783ed 100644
--- a/parser/src/parser/context.rs
+++ b/parser/src/parser/context.rs
@@ -153,7 +153,7 @@ impl ParserContext {
                 }
                 _ => GroupCoordinate::SignLow,
             },
-            Token::Ident(x) if x == "_" => GroupCoordinate::Inferred,
+            Token::Underscore => GroupCoordinate::Inferred,
             Token::Int(value) => GroupCoordinate::Number(value.clone(), token.span.clone()),
             _ => return None,
         })
diff --git a/parser/src/tokenizer/lexer.rs b/parser/src/tokenizer/lexer.rs
index d079a807a8..d3b04e54d9 100644
--- a/parser/src/tokenizer/lexer.rs
+++ b/parser/src/tokenizer/lexer.rs
@@ -43,14 +43,11 @@ fn eat_identifier(input: &[u8]) -> Option<(&[u8], &[u8])> {
     if input.is_empty() {
         return None;
     }
-    if !input[0].is_ascii_alphabetic() && input[0] != b'_' {
+    if !input[0].is_ascii_alphabetic() {
         // Allow _ at start.
         return None;
     }
-    if input.len() == 1 && input[0] == b'_' {
-        // But only if it the length of the identifer is 1.
-        return None;
-    }
+
     let mut i = 1usize;
     while i < input.len() {
         if !input[i].is_ascii_alphanumeric() && input[i] != b'_' {
@@ -283,6 +280,7 @@ impl Token {
                 }
                 return (&input[1..], Some(Token::Assign));
             }
+            b'_' => return (&input[1..], Some(Token::Underscore)),
             b'@' => return (&input[1..], Some(Token::At)),
             b'[' => return (&input[1..], Some(Token::LeftSquare)),
             b']' => return (&input[1..], Some(Token::RightSquare)),
diff --git a/parser/src/tokenizer/mod.rs b/parser/src/tokenizer/mod.rs
index 7cfe52175f..1d4848f7d4 100644
--- a/parser/src/tokenizer/mod.rs
+++ b/parser/src/tokenizer/mod.rs
@@ -172,6 +172,7 @@ mod tests {
         -
         -=
         ->
+        _
         .
         ..
         ...
@@ -221,7 +222,7 @@ mod tests {
         }
         assert_eq!(
             output,
-            r#""test" "test{}test" "test{}" "{}test" "test{" "test}" "test{test" "test}test" "te{{}}" aleo1qnr4dkkvkgfqph0vzc3y6z2eu975wnpz2925ntjccd5cfqxtyu8sta57j8 test_ident 12345 address as bool circuit const else false field for function group i128 i64 i32 i16 i8 if import in input let mut return static string test true u128 u64 u32 u16 u8 self Self console ! != && ( ) * ** **= *= + += , - -= -> . .. ... / /= : :: ; < <= = == > >= @ [ ] { { } } || & &= | |= ^ ^= ~ << <<= >> >>= >>> >>>= % %= ||= &&= ? // test
+            r#""test" "test{}test" "test{}" "{}test" "test{" "test}" "test{test" "test}test" "te{{}}" aleo1qnr4dkkvkgfqph0vzc3y6z2eu975wnpz2925ntjccd5cfqxtyu8sta57j8 test_ident 12345 address as bool circuit const else false field for function group i128 i64 i32 i16 i8 if import in input let mut return static string test true u128 u64 u32 u16 u8 self Self console ! != && ( ) * ** **= *= + += , - -= -> _ . .. ... / /= : :: ; < <= = == > >= @ [ ] { { } } || & &= | |= ^ ^= ~ << <<= >> >>= >>> >>>= % %= ||= &&= ? // test
  /* test */ //
  "#
         );
diff --git a/parser/src/tokenizer/token.rs b/parser/src/tokenizer/token.rs
index 43e6250bc7..3af070fe26 100644
--- a/parser/src/tokenizer/token.rs
+++ b/parser/src/tokenizer/token.rs
@@ -85,6 +85,7 @@ pub enum Token {
     DoubleColon,
     Question,
     Arrow,
+    Underscore,
 
     // Syntactic Grammr
     // Types
@@ -298,6 +299,7 @@ impl fmt::Display for Token {
             OrEq => write!(f, "||="),
             AndEq => write!(f, "&&="),
             Question => write!(f, "?"),
+            Underscore => write!(f, "_"),
         }
     }
 }