diff --git a/vllm/v1/core/sched/scheduler.py b/vllm/v1/core/sched/scheduler.py
index 9a2a589cfae4..d57ccc400b88 100644
--- a/vllm/v1/core/sched/scheduler.py
+++ b/vllm/v1/core/sched/scheduler.py
@@ -1676,12 +1676,13 @@ def _update_waiting_for_remote_kv(self, request: Request) -> bool:
             num_computed_tokens = len(block_ids) * self.block_size
             # Handle the case where num request tokens less than one block.
             num_computed_tokens = min(num_computed_tokens, request.num_tokens)
-            if num_computed_tokens == request.num_tokens:
-                num_computed_tokens -= 1
             # This will cache the blocks iff caching is enabled.
             self.kv_cache_manager.cache_blocks(request, num_computed_tokens)
 
             # Update the request state for scheduling.
+            if num_computed_tokens == request.num_tokens:
+                # need to recompute last token in order to sample the next token
+                num_computed_tokens -= 1
             request.num_computed_tokens = num_computed_tokens
 
         # Return that we are ready.