diff --git a/doc/source/whatsnew/v3.0.0.rst b/doc/source/whatsnew/v3.0.0.rst
index c2a56afbc580e..3b5183c43bcd0 100644
--- a/doc/source/whatsnew/v3.0.0.rst
+++ b/doc/source/whatsnew/v3.0.0.rst
@@ -630,6 +630,7 @@ I/O
 - Bug in :meth:`read_csv` where the order of the ``na_values`` makes an inconsistency when ``na_values`` is a list non-string values. (:issue:`59303`)
 - Bug in :meth:`read_excel` raising ``ValueError`` when passing array of boolean values when ``dtype="boolean"``. (:issue:`58159`)
 - Bug in :meth:`read_json` not validating the ``typ`` argument to not be exactly ``"frame"`` or ``"series"`` (:issue:`59124`)
+- Bug in :meth:`read_json` where extreme value integers in string format were incorrectly parsed as a different integer number (:issue:`20608`)
 - Bug in :meth:`read_stata` raising ``KeyError`` when input file is stored in big-endian format and contains strL data. (:issue:`58638`)
 - Bug in :meth:`read_stata` where extreme value integers were incorrectly interpreted as missing for format versions 111 and prior (:issue:`58130`)
 - Bug in :meth:`read_stata` where the missing code for double was not recognised for format versions 105 and prior (:issue:`58149`)
diff --git a/pandas/io/json/_json.py b/pandas/io/json/_json.py
index d077b9e0c4568..e9c9f5ba225a5 100644
--- a/pandas/io/json/_json.py
+++ b/pandas/io/json/_json.py
@@ -1168,6 +1168,7 @@ def _try_convert_data(
         """
         Try to parse a Series into a column by inferring dtype.
         """
+        org_data = data
         # don't try to coerce, unless a force conversion
         if use_dtypes:
             if not self.dtype:
@@ -1222,7 +1223,7 @@ def _try_convert_data(
         if len(data) and data.dtype in ("float", "object"):
             # coerce ints if we can
             try:
-                new_data = data.astype("int64")
+                new_data = org_data.astype("int64")
                 if (new_data == data).all():
                     data = new_data
                     converted = True
diff --git a/pandas/tests/io/json/test_pandas.py b/pandas/tests/io/json/test_pandas.py
index 1c54232b8b510..d3328d1dfcaef 100644
--- a/pandas/tests/io/json/test_pandas.py
+++ b/pandas/tests/io/json/test_pandas.py
@@ -2286,3 +2286,15 @@ def test_read_json_lines_rangeindex():
     result = read_json(StringIO(data), lines=True).index
     expected = RangeIndex(2)
     tm.assert_index_equal(result, expected, exact=True)
+
+
+def test_large_number():
+    # GH#20608
+    result = read_json(
+        StringIO('["9999999999999999"]'),
+        orient="values",
+        typ="series",
+        convert_dates=False,
+    )
+    expected = Series([9999999999999999])
+    tm.assert_series_equal(result, expected)