diff --git a/pymc/logprob/mixture.py b/pymc/logprob/mixture.py
index de1d5a9a16..b9bf2283c1 100644
--- a/pymc/logprob/mixture.py
+++ b/pymc/logprob/mixture.py
@@ -445,6 +445,10 @@ def logprob_MixtureRV(
             logp_val = at.set_subtensor(logp_val[idx_m_on_axis], logp_m)
 
     else:
+        # FIXME: This logprob implementation does not support mixing across distinct components,
+        # but we sometimes use it, because MixtureRV does not keep information about at which
+        # dimension scalar indexing actually starts
+
         # If the stacking operation expands the component RVs, we have
         # to expand the value and later squeeze the logprob for everything
         # to work correctly
diff --git a/pymc/tests/logprob/test_mixture.py b/pymc/tests/logprob/test_mixture.py
index 8deff51982..683e6b2fa1 100644
--- a/pymc/tests/logprob/test_mixture.py
+++ b/pymc/tests/logprob/test_mixture.py
@@ -306,6 +306,31 @@ def test_hetero_mixture_binomial(p_val, size):
             (slice(None),),
             1,
         ),
+        # Vector mixture components, scalar index that mixes across components
+        pytest.param(
+            (
+                np.array(0, dtype=pytensor.config.floatX),
+                np.array(1, dtype=pytensor.config.floatX),
+            ),
+            (
+                np.array(0.5, dtype=pytensor.config.floatX),
+                np.array(0.5, dtype=pytensor.config.floatX),
+            ),
+            (
+                np.array(100, dtype=pytensor.config.floatX),
+                np.array(1, dtype=pytensor.config.floatX),
+            ),
+            np.array([0.1, 0.5, 0.1, 0.3], dtype=pytensor.config.floatX),
+            (4,),
+            (),
+            (),
+            1,
+            marks=pytest.mark.xfail(
+                AssertionError,
+                match="Arrays are not almost equal to 6 decimals",  # This is ignored, but that's where it should fail!
+                reason="IfElse Mixture logprob fails when indexing mixes across components",
+            ),
+        ),
         # Matrix components, scalar index along first axis
         (
             (