diff --git a/ch05/11_qwen3/qwen3-chat-interface/qwen3-chat-interface-multiturn.py b/ch05/11_qwen3/qwen3-chat-interface/qwen3-chat-interface-multiturn.py
index be5f3ba..37a4366 100644
--- a/ch05/11_qwen3/qwen3-chat-interface/qwen3-chat-interface-multiturn.py
+++ b/ch05/11_qwen3/qwen3-chat-interface/qwen3-chat-interface-multiturn.py
@@ -115,6 +115,10 @@ REPO_ID, LOCAL_DIR = build_repo_and_local(MODEL, REASONING, LOCAL_DIR)
 DEVICE = get_device(DEVICE)
 MODEL, TOKENIZER = get_model_and_tokenizer(QWEN3_CONFIG, REPO_ID, LOCAL_DIR, DEVICE, REASONING)
 
+# Even though the official TOKENIZER.eos_token_id is either <|im_end|> (reasoning)
+# or <|endoftext|> (base), the reasoning model sometimes emits both.
+EOS_TOKEN_IDS = (TOKENIZER.encode("<|im_end|>")[0], TOKENIZER.encode("<|endoftext|>")[0])
+
 
 @chainlit.on_chat_start
 async def on_start():
@@ -147,9 +151,11 @@ async def main(message: chainlit.Message):
         model=MODEL,
         token_ids=input_ids_tensor,
         max_new_tokens=MAX_NEW_TOKENS,
-        eos_token_id=TOKENIZER.eos_token_id
+        # eos_token_id=TOKENIZER.eos_token_id
     ):
         token_id = tok.squeeze(0)
+        if token_id in EOS_TOKEN_IDS:
+            break
         piece = TOKENIZER.decode(token_id.tolist())
         await out_msg.stream_token(piece)