diff --git a/LLama.Examples/Examples/BatchedDecoding.cs b/LLama.Examples/Examples/BatchedDecoding.cs
index 1d55ff12..c51c49be 100644
--- a/LLama.Examples/Examples/BatchedDecoding.cs
+++ b/LLama.Examples/Examples/BatchedDecoding.cs
@@ -52,7 +52,7 @@ public class BatchedDecoding
             return;
         }
 
-        var batch = new LLamaBatch(1);
+        var batch = new LLamaBatch();
 
         // evaluate the initial prompt
         for (var i = 0; i < prompt_tokens.Length; i++)
diff --git a/LLama/Native/LLamaBatch.cs b/LLama/Native/LLamaBatch.cs
index a7ca923b..20e14530 100644
--- a/LLama/Native/LLamaBatch.cs
+++ b/LLama/Native/LLamaBatch.cs
@@ -38,7 +38,7 @@ public class LLamaBatch
     {
         // These can both be grown later, start off with reasonable numbers.
         const int n_tokens = 128;
-        const int n_seq_max = 4;
+        const int n_seq_max = 1;
 
         MaxSequences = n_seq_max;
         TokenCapacity = n_tokens;
@@ -77,9 +77,9 @@ public class LLamaBatch
         }
     }
 
-    private void GrowMaxSequences()
+    private void GrowMaxSequences(int atLeast)
     {
-        var n_seq = MaxSequences * 2;
+        var n_seq = Math.Max(MaxSequences * 2, atLeast);
         MaxSequences = n_seq;
 
         for (var i = 0; i < _sequenceIds.Length; i++)
@@ -130,7 +130,7 @@ public class LLamaBatch
         if (TokenCount == TokenCapacity)
             GrowTokenCapacity();
         if (sequences.Length > MaxSequences)
-            GrowMaxSequences();
+            GrowMaxSequences(sequences.Length);
 
         _tokens[TokenCount] = token;
         _positions[TokenCount] = pos;