diff --git a/examples/benchmark.ipynb b/examples/benchmark.ipynb
index 1ff00d4..758b19a 100644
--- a/examples/benchmark.ipynb
+++ b/examples/benchmark.ipynb
@@ -253,7 +253,7 @@
     "\n",
     "            if is_vanilla_llm:\n",
     "                llm = agent\n",
-    "                answer = str(llm([{\"role\": \"user\", \"content\": question}]))\n",
+    "                answer = str(llm([{\"role\": \"user\", \"content\": question}]).content)\n",
     "                token_count = {\"input\": llm.last_input_token_count, \"output\": llm.last_output_token_count}\n",
     "                intermediate_steps = str([])\n",
     "            else:\n",