diff --git a/vllm/v1/worker/block_table.py b/vllm/v1/worker/block_table.py
index 0f5446b44..f46e8a8ed 100644
--- a/vllm/v1/worker/block_table.py
+++ b/vllm/v1/worker/block_table.py
@@ -121,6 +121,12 @@ class BlockTable:
         self.num_blocks_per_row[row_idx] = 0
         self.append_row(block_ids, row_idx)
 
+    def clear_row(self, row_idx: int) -> None:
+        num_blocks = self.num_blocks_per_row[row_idx]
+        if num_blocks > 0:
+            self.block_table.np[row_idx, :num_blocks] = 0
+        self.num_blocks_per_row[row_idx] = 0
+
     def move_row(self, src: int, tgt: int) -> None:
         num_blocks = self.num_blocks_per_row[src]
         block_table_np = self.block_table.np
@@ -275,6 +281,10 @@ class MultiGroupBlockTable:
         for i, block_table in enumerate(self.block_tables):
             block_table.add_row(block_ids[i], row_idx)
 
+    def clear_row(self, row_idx: int) -> None:
+        for block_table in self.block_tables:
+            block_table.clear_row(row_idx)
+
     def move_row(self, src: int, tgt: int) -> None:
         for block_table in self.block_tables:
             block_table.move_row(src, tgt)
diff --git a/vllm/v1/worker/gpu_input_batch.py b/vllm/v1/worker/gpu_input_batch.py
index e20d268fe..11d57f1d7 100644
--- a/vllm/v1/worker/gpu_input_batch.py
+++ b/vllm/v1/worker/gpu_input_batch.py
@@ -496,6 +496,7 @@ class InputBatch:
         self._req_ids[req_index] = None
         self.req_output_token_ids[req_index] = None
         self.spec_token_ids[req_index].clear()
+        self.block_table.clear_row(req_index)
 
         # LoRA
         lora_id = self.request_lora_mapping[req_index]
diff --git a/vllm/v1/worker/gpu_model_runner.py b/vllm/v1/worker/gpu_model_runner.py
index 6465ca654..be7734487 100644
--- a/vllm/v1/worker/gpu_model_runner.py
+++ b/vllm/v1/worker/gpu_model_runner.py
@@ -5376,6 +5376,12 @@ class GPUModelRunner(
                 self.query_start_loc.np[1 : num_reqs + 1] = cum_num_tokens
                 self.query_start_loc.copy_to_gpu()
 
+                # Sync block table CPU->GPU so cleared rows from
+                # remove_request() are visible to the attention metadata
+                # builder. Without this, stale block IDs from finished
+                # requests can corrupt Mamba state.
+                self.input_batch.block_table.commit_block_table(num_reqs_padded)
+
                 pad_attn = cudagraph_runtime_mode == CUDAGraphMode.FULL
                 attn_metadata, _ = self._build_attention_metadata(
                     num_tokens=num_tokens_unpadded,