diff --git a/Userland/Libraries/LibJS/Lexer.cpp b/Userland/Libraries/LibJS/Lexer.cpp
index 340eca7ec5..cd4dc8a415 100644
--- a/Userland/Libraries/LibJS/Lexer.cpp
+++ b/Userland/Libraries/LibJS/Lexer.cpp
@@ -27,6 +27,7 @@ Lexer::Lexer(StringView source, StringView filename, size_t line_number, size_t
     , m_filename(filename)
     , m_line_number(line_number)
     , m_line_column(line_column)
+    , m_parsed_identifiers(adopt_ref(*new ParsedIdentifiers))
 {
     if (s_keywords.is_empty()) {
         s_keywords.set("await", TokenType::Await);
@@ -602,8 +603,7 @@ Token Lexer::next()
         } while (code_point.has_value());
 
         identifier = builder.build();
-        if (!m_parsed_identifiers.contains_slow(*identifier))
-            m_parsed_identifiers.append(*identifier);
+        m_parsed_identifiers->identifiers.set(*identifier);
 
         auto it = s_keywords.find(identifier->hash(), [&](auto& entry) { return entry.key == identifier; });
         if (it == s_keywords.end())
diff --git a/Userland/Libraries/LibJS/Lexer.h b/Userland/Libraries/LibJS/Lexer.h
index 03991f596d..1ecbac223b 100644
--- a/Userland/Libraries/LibJS/Lexer.h
+++ b/Userland/Libraries/LibJS/Lexer.h
@@ -82,9 +82,13 @@ private:
     static HashMap<String, TokenType> s_two_char_tokens;
     static HashMap<char, TokenType> s_single_char_tokens;
 
-    // Resolved identifiers must be kept alive for the duration of the parsing stage, otherwise
-    // the only references to these strings are deleted by the Token destructor.
-    Vector<FlyString> m_parsed_identifiers;
+    struct ParsedIdentifiers : public RefCounted<ParsedIdentifiers> {
+        // Resolved identifiers must be kept alive for the duration of the parsing stage, otherwise
+        // the only references to these strings are deleted by the Token destructor.
+        HashTable<FlyString> identifiers;
+    };
+
+    RefPtr<ParsedIdentifiers> m_parsed_identifiers;
 };
 
 }