diff --git a/src/Lexer.ts b/src/Lexer.ts
index d600922bd7..9aa79cdbf9 100644
--- a/src/Lexer.ts
+++ b/src/Lexer.ts
@@ -106,10 +106,6 @@ export class _Lexer {
   blockTokens(src: string, tokens: Token[] = [], lastParagraphClipped = false) {
     if (this.options.pedantic) {
       src = src.replace(/\t/g, '    ').replace(/^ +$/gm, '');
-    } else {
-      src = src.replace(/^( *)(\t+)/gm, (_, leading, tabs) => {
-        return leading + '    '.repeat(tabs.length);
-      });
     }
 
     let token: Tokens.Generic | undefined;
diff --git a/src/Tokenizer.ts b/src/Tokenizer.ts
index 5ad91c61e1..80fc59ec7a 100644
--- a/src/Tokenizer.ts
+++ b/src/Tokenizer.ts
@@ -90,7 +90,7 @@ export class _Tokenizer {
   code(src: string): Tokens.Code | undefined {
     const cap = this.rules.block.code.exec(src);
     if (cap) {
-      const text = cap[0].replace(/^ {1,4}/gm, '');
+      const text = cap[0].replace(/^(?: {1,4}| {0,3}\t)/gm, '');
       return {
         type: 'code',
         raw: cap[0],
@@ -294,7 +294,7 @@ export class _Tokenizer {
           indent += cap[1].length;
         }
 
-        if (blankLine && /^ *$/.test(nextLine)) { // Items begin with at most one blank line
+        if (blankLine && /^[ \t]*$/.test(nextLine)) { // Items begin with at most one blank line
           raw += nextLine + '\n';
           src = src.substring(nextLine.length + 1);
           endEarly = true;
@@ -309,11 +309,15 @@ export class _Tokenizer {
           // Check if following lines should be included in List Item
           while (src) {
             const rawLine = src.split('\n', 1)[0];
+            let nextLineWithoutTabs;
             nextLine = rawLine;
 
             // Re-align to follow commonmark nesting rules
             if (this.options.pedantic) {
               nextLine = nextLine.replace(/^ {1,4}(?=( {4})*[^ ])/g, '  ');
+              nextLineWithoutTabs = nextLine;
+            } else {
+              nextLineWithoutTabs = nextLine.replace(/\t/g, '    ');
             }
 
             // End list item if found code fences
@@ -332,12 +336,12 @@ export class _Tokenizer {
             }
 
             // Horizontal rule found
-            if (hrRegex.test(src)) {
+            if (hrRegex.test(nextLine)) {
               break;
             }
 
-            if (nextLine.search(/[^ ]/) >= indent || !nextLine.trim()) { // Dedent if possible
-              itemContents += '\n' + nextLine.slice(indent);
+            if (nextLineWithoutTabs.search(/[^ ]/) >= indent || !nextLine.trim()) { // Dedent if possible
+              itemContents += '\n' + nextLineWithoutTabs.slice(indent);
             } else {
               // not enough indentation
               if (blankLine) {
@@ -345,7 +349,7 @@ export class _Tokenizer {
               }
 
               // paragraph continuation unless last line was a different block level element
-              if (line.search(/[^ ]/) >= 4) { // indented code block
+              if (line.replace(/\t/g, '    ').search(/[^ ]/) >= 4) { // indented code block
                 break;
               }
               if (fencesBeginRegex.test(line)) {
@@ -367,7 +371,7 @@ export class _Tokenizer {
 
             raw += rawLine + '\n';
             src = src.substring(rawLine.length + 1);
-            line = nextLine.slice(indent);
+            line = nextLineWithoutTabs.slice(indent);
           }
         }
 
@@ -375,7 +379,7 @@ export class _Tokenizer {
           // If the previous item ended with a blank line, the list is loose
           if (endsWithBlankLine) {
             list.loose = true;
-          } else if (/\n *\n *$/.test(raw)) {
+          } else if (/\n[ \t]*\n[ \t]*$/.test(raw)) {
             endsWithBlankLine = true;
           }
         }
diff --git a/src/rules.ts b/src/rules.ts
index d6f330ed39..83bc8b3f1f 100644
--- a/src/rules.ts
+++ b/src/rules.ts
@@ -6,15 +6,15 @@ import {
  * Block-Level Grammar
  */
 
-const newline = /^(?: *(?:\n|$))+/;
-const blockCode = /^( {4}[^\n]+(?:\n(?: *(?:\n|$))*)?)+/;
+const newline = /^(?:[ \t]*(?:\n|$))+/;
+const blockCode = /^((?: {4}| {0,3}\t)[^\n]+(?:\n(?:[ \t]*(?:\n|$))*)?)+/;
 const fences = /^ {0,3}(`{3,}(?=[^`\n]*(?:\n|$))|~{3,})([^\n]*)(?:\n|$)(?:|([\s\S]*?)(?:\n|$))(?: {0,3}\1[~`]* *(?=\n|$)|$)/;
 const hr = /^ {0,3}((?:-[\t ]*){3,}|(?:_[ \t]*){3,}|(?:\*[ \t]*){3,})(?:\n+|$)/;
 const heading = /^ {0,3}(#{1,6})(?=\s|$)(.*)(?:\n+|$)/;
 const bullet = /(?:[*+-]|\d{1,9}[.)])/;
 const lheading = edit(/^(?!bull |blockCode|fences|blockquote|heading|html)((?:.|\n(?!\s*?\n|bull |blockCode|fences|blockquote|heading|html))+?)\n {0,3}(=+|-+) *(?:\n+|$)/)
   .replace(/bull/g, bullet) // lists can interrupt
-  .replace(/blockCode/g, / {4}/) // indented code blocks can interrupt
+  .replace(/blockCode/g, /(?: {4}| {0,3}\t)/) // indented code blocks can interrupt
   .replace(/fences/g, / {0,3}(?:`{3,}|~{3,})/) // fenced code blocks can interrupt
   .replace(/blockquote/g, / {0,3}>/) // blockquote can interrupt
   .replace(/heading/g, / {0,3}#{1,6}/) // ATX heading can interrupt
@@ -23,7 +23,7 @@ const lheading = edit(/^(?!bull |blockCode|fences|blockquote|heading|html)((?:.|
 const _paragraph = /^([^\n]+(?:\n(?!hr|heading|lheading|blockquote|fences|list|html|table| +\n)[^\n]+)*)/;
 const blockText = /^[^\n]+/;
 const _blockLabel = /(?!\s*\])(?:\\.|[^\[\]\\])+/;
-const def = edit(/^ {0,3}\[(label)\]: *(?:\n *)?([^<\s][^\s]*|<.*?>)(?:(?: +(?:\n *)?| *\n *)(title))? *(?:\n+|$)/)
+const def = edit(/^ {0,3}\[(label)\]: *(?:\n[ \t]*)?([^<\s][^\s]*|<.*?>)(?:(?: +(?:\n[ \t]*)?| *\n[ \t]*)(title))? *(?:\n+|$)/)
   .replace('label', _blockLabel)
   .replace('title', /(?:"(?:\\"?|[^"\\])*"|'[^'\n]*(?:\n[^'\n]+)*\n?'|\([^()]*\))/)
   .getRegex();
@@ -46,9 +46,9 @@ const html = edit(
 + '|<\\?[\\s\\S]*?(?:\\?>\\n*|$)' // (3)
 + '|<![A-Z][\\s\\S]*?(?:>\\n*|$)' // (4)
 + '|<!\\[CDATA\\[[\\s\\S]*?(?:\\]\\]>\\n*|$)' // (5)
-+ '|</?(tag)(?: +|\\n|/?>)[\\s\\S]*?(?:(?:\\n *)+\\n|$)' // (6)
-+ '|<(?!script|pre|style|textarea)([a-z][\\w-]*)(?:attribute)*? */?>(?=[ \\t]*(?:\\n|$))[\\s\\S]*?(?:(?:\\n *)+\\n|$)' // (7) open tag
-+ '|</(?!script|pre|style|textarea)[a-z][\\w-]*\\s*>(?=[ \\t]*(?:\\n|$))[\\s\\S]*?(?:(?:\\n *)+\\n|$)' // (7) closing tag
++ '|</?(tag)(?: +|\\n|/?>)[\\s\\S]*?(?:(?:\\n[ \t]*)+\\n|$)' // (6)
++ '|<(?!script|pre|style|textarea)([a-z][\\w-]*)(?:attribute)*? */?>(?=[ \\t]*(?:\\n|$))[\\s\\S]*?(?:(?:\\n[ \t]*)+\\n|$)' // (7) open tag
++ '|</(?!script|pre|style|textarea)[a-z][\\w-]*\\s*>(?=[ \\t]*(?:\\n|$))[\\s\\S]*?(?:(?:\\n[ \t]*)+\\n|$)' // (7) closing tag
 + ')', 'i')
   .replace('comment', _comment)
   .replace('tag', _tag)
@@ -104,7 +104,7 @@ const gfmTable = edit(
   .replace('hr', hr)
   .replace('heading', ' {0,3}#{1,6}(?:\\s|$)')
   .replace('blockquote', ' {0,3}>')
-  .replace('code', ' {4}[^\\n]')
+  .replace('code', '(?: {4}| {0,3}\t)[^\\n]')
   .replace('fences', ' {0,3}(?:`{3,}(?=[^`\\n]*\\n)|~{3,})[^\\n]*\\n')
   .replace('list', ' {0,3}(?:[*+-]|1[.)]) ') // only lists starting from 1 can interrupt
   .replace('html', '</?(?:tag)(?: +|\\n|/?>)|<(?:script|pre|style|textarea|!--)')
diff --git a/test/specs/new/tabs_code.html b/test/specs/new/tabs_code.html
new file mode 100644
index 0000000000..f8752e19f9
--- /dev/null
+++ b/test/specs/new/tabs_code.html
@@ -0,0 +1,2 @@
+<pre><code>	tab
+</code></pre>
diff --git a/test/specs/new/tabs_code.md b/test/specs/new/tabs_code.md
new file mode 100644
index 0000000000..1689e98aa1
--- /dev/null
+++ b/test/specs/new/tabs_code.md
@@ -0,0 +1,6 @@
+---
+renderExact: true
+---
+```
+	tab
+```