diff --git a/src/corner/core.py b/src/corner/core.py
index 0b160d2..a518425 100644
--- a/src/corner/core.py
+++ b/src/corner/core.py
@@ -526,7 +526,7 @@ def quantile(x, q, weights=None):
             raise ValueError("Dimension mismatch: len(weights) != len(x)")
         idx = np.argsort(x)
         sw = weights[idx]
-        cdf = np.cumsum(sw)[:-1]
+        cdf = np.cumsum(sw, dtype=np.result_type(sw.dtype, np.float64))[:-1]
         cdf /= cdf[-1]
         cdf = np.append(0, cdf)
         return np.interp(q, cdf, x[idx]).tolist()
diff --git a/tests/test_quantile.py b/tests/test_quantile.py
index e854204..d761122 100644
--- a/tests/test_quantile.py
+++ b/tests/test_quantile.py
@@ -53,3 +53,22 @@ def test_weighted_quantile(seed=42):
     q = [0.0, 1.0]
     a = corner.quantile(x, q, weights=np.random.rand(len(x)))
     assert np.allclose(a, (np.min(x), np.max(x)))
+
+
+@pytest.mark.parametrize("dtype", [np.int8, np.int32, np.uint64])
+def test_integer_weights(dtype):
+    x = np.array([7.0, 2.0, 9.0, 1.0, 5.0])
+    weights = np.array([90, 40, 30, 100, 60], dtype=dtype)
+    q = [0.0, 0.16, 0.5, 0.84, 1.0]
+    actual = corner.quantile(x, q, weights=weights)
+    expected = corner.quantile(x, q, weights=weights.astype(float))
+    np.testing.assert_array_equal(actual, expected)
+
+
+def test_python_integer_weights():
+    x = np.arange(10.0)
+    q = [0.16, 0.5, 0.84]
+    np.testing.assert_allclose(
+        corner.quantile(x, q, weights=[1] * 10),
+        np.percentile(x, 100 * np.array(q)),
+    )