diff --git a/vllm/model_executor/layers/sampler.py b/vllm/model_executor/layers/sampler.py
index 13da9aa38af0..adb6d9b1f6ae 100644
--- a/vllm/model_executor/layers/sampler.py
+++ b/vllm/model_executor/layers/sampler.py
@@ -539,10 +539,13 @@ def _get_logprobs(
             prompt_len = sampling_metadata.prompt_lens[i]
             prompt_tokens = sampling_metadata.seq_data[
                 seq_ids[0]].prompt_token_ids
+            # Swapped seqs have output tokens.
+            output_tokens = sampling_metadata.seq_data[
+                seq_ids[0]].output_token_ids
             batched_logprobs_query_seq_indices.extend(
                 sample_idx + j for j in range(prompt_len - 1))
             batched_logprobs_query_token_indices.extend(
-                token_id for token_id in prompt_tokens[1:])
+                token_id for token_id in prompt_tokens[1:] + output_tokens)
             sample_idx += prompt_len - 1
         batched_logprobs_query_seq_indices.extend(
             [sample_idx + parent_id for parent_id in parent_ids])
@@ -586,8 +589,11 @@ def _get_logprobs(
             prompt_len = sampling_metadata.prompt_lens[i]
             prompt_tokens = sampling_metadata.seq_data[
                 seq_ids[0]].prompt_token_ids
+            # Swapped seqs have output tokens.
+            output_tokens = sampling_metadata.seq_data[
+                seq_ids[0]].output_token_ids
             group_prompt_logprobs: PromptLogprobs = [None]
-            for token_id in prompt_tokens[1:]:
+            for token_id in prompt_tokens[1:] + output_tokens:
                 prompt_logprobs_dict = {
                     token_id:
                     batched_logprobs_query_result[query_result_idx].item()