diff --git a/python/sglang/srt/layers/attention/flashinfer_backend.py b/python/sglang/srt/layers/attention/flashinfer_backend.py
index e899bcb7e..99708135a 100644
--- a/python/sglang/srt/layers/attention/flashinfer_backend.py
+++ b/python/sglang/srt/layers/attention/flashinfer_backend.py
@@ -947,7 +947,7 @@ class FlashInferMultiStepDraftBackend:
             triton.next_power_of_2(bs),
         )
 
-        for i in range(self.speculative_num_steps):
+        for i in range(self.speculative_num_steps - 1):
             forward_batch.spec_info.kv_indptr = self.kv_indptr[i, : bs + 1]
             forward_batch.spec_info.kv_indices = kv_indices_buffer[i][
                 : seq_lens_sum * self.topk + bs * (i + 1)
diff --git a/python/sglang/srt/speculative/eagle_worker.py b/python/sglang/srt/speculative/eagle_worker.py
index 45798ba58..8c24f2aa5 100644
--- a/python/sglang/srt/speculative/eagle_worker.py
+++ b/python/sglang/srt/speculative/eagle_worker.py
@@ -234,6 +234,10 @@ class EAGLEWorker(TpModelWorker):
             token_list.append(tree_info[1])
             parents_list.append(tree_info[2])
 
+            # we don't need to run the last forward. we get 1 token from draft prefill and (#spec steps - 1) tokens here
+            if i == self.speculative_num_steps - 1:
+                break
+
             # Set inputs
             forward_batch.input_ids = input_ids
             forward_batch.out_cache_loc = out_cache_loc[