diff --git a/botorch/optim/optimize.py b/botorch/optim/optimize.py
index 6f3a5876a9..ef37c38d87 100644
--- a/botorch/optim/optimize.py
+++ b/botorch/optim/optimize.py
@@ -196,6 +196,8 @@ def _optimize_acqf_all_features_fixed(
     X = X.expand(q, *X.shape)
     with torch.no_grad():
         acq_value = acq_function(X)
+    if acq_value.ndim == 1:
+        acq_value = acq_value[0]
     return X, acq_value
 
 
diff --git a/test/optim/test_optimize.py b/test/optim/test_optimize.py
index 8d8be47ea0..e0bff2ce62 100644
--- a/test/optim/test_optimize.py
+++ b/test/optim/test_optimize.py
@@ -14,10 +14,12 @@
 
 import numpy as np
 import torch
+
 from botorch.acquisition.acquisition import (
     AcquisitionFunction,
     OneShotAcquisitionFunction,
 )
+from botorch.acquisition.analytic import LogExpectedImprovement
 from botorch.acquisition.knowledge_gradient import qKnowledgeGradient
 from botorch.acquisition.monte_carlo import qExpectedImprovement
 from botorch.acquisition.multi_objective.hypervolume_knowledge_gradient import (
@@ -1147,6 +1149,23 @@ def nlc(x):
                 ),
             )
 
+    def test_optimize_acqf_all_fixed_features(self):
+        train_X = torch.rand(3, 2)
+        train_Y = torch.rand(3, 1)
+        gp = SingleTaskGP(train_X=train_X, train_Y=train_Y)
+        gp.eval()
+        logEI = LogExpectedImprovement(model=gp, best_f=train_Y.max())
+        bounds = torch.stack([torch.zeros(2), torch.ones(2)])
+        _, acqf_value = optimize_acqf(
+            logEI,
+            bounds,
+            q=1,
+            num_restarts=1,
+            raw_samples=1,
+            fixed_features={0: 0, 1: 0},
+        )
+        self.assertEqual(acqf_value.ndim, 0)
+
     def test_constraint_caching(self):
         def nlc(x):
             return 4 - x.sum(dim=-1)