diff --git a/configs/datasets/truthfulqa/truthfulqa_gen_1e7d8d.py b/configs/datasets/truthfulqa/truthfulqa_gen_1e7d8d.py
index 269ab9486..74fa449c3 100644
--- a/configs/datasets/truthfulqa/truthfulqa_gen_1e7d8d.py
+++ b/configs/datasets/truthfulqa/truthfulqa_gen_1e7d8d.py
@@ -18,8 +18,11 @@
 # Metrics such as 'truth' and 'info' needs
 # OPENAI_API_KEY with finetuned models in it.
 # Please use your own finetuned openai model with keys and refers to
-# the source code for more details
-# Metrics such as 'bleurt', 'rouge', 'bleu' are free to test
+# the source code of `TruthfulQAEvaluator` for more details.
+#
+# If you cannot provide available models for 'truth' and 'info',
+# and want to perform basic metric eval, please set
+# `metrics=('bleurt', 'rouge', 'bleu')`
 
 # When key is set to "ENV", the key will be fetched from the environment
 # variable $OPENAI_API_KEY. Otherwise, set key in here directly.
diff --git a/configs/datasets/truthfulqa/truthfulqa_gen_5ddc62.py b/configs/datasets/truthfulqa/truthfulqa_gen_5ddc62.py
index 47227cc63..804675409 100644
--- a/configs/datasets/truthfulqa/truthfulqa_gen_5ddc62.py
+++ b/configs/datasets/truthfulqa/truthfulqa_gen_5ddc62.py
@@ -20,8 +20,11 @@
 # Metrics such as 'truth' and 'info' needs
 # OPENAI_API_KEY with finetuned models in it.
 # Please use your own finetuned openai model with keys and refers to
-# the source code for more details
-# Metrics such as 'bleurt', 'rouge', 'bleu' are free to test
+# the source code of `TruthfulQAEvaluator` for more details.
+#
+# If you cannot provide available models for 'truth' and 'info',
+# and want to perform basic metric eval, please set
+# `metrics=('bleurt', 'rouge', 'bleu')`
 
 # When key is set to "ENV", the key will be fetched from the environment
 # variable $OPENAI_API_KEY. Otherwise, set key in here directly.
diff --git a/opencompass/datasets/truthfulqa.py b/opencompass/datasets/truthfulqa.py
index a81cfc446..8ee46af3f 100644
--- a/opencompass/datasets/truthfulqa.py
+++ b/opencompass/datasets/truthfulqa.py
@@ -39,7 +39,9 @@ class TruthfulQAEvaluator(BaseEvaluator):
 
     Args:
         truth_model (str): Truth model name. See "notes" for details.
+            Defaults to ''.
         info_model (str): Informativeness model name. See "notes" for details.
+            Defaults to ''.
         metrics (tuple): Computing needed metrics for truthfulqa dataset.
             Supported metrics are `bleurt`, `rouge`, `bleu`, `truth`, `info`.
         key (str): Corresponding API key. If set to `ENV`, find it in
@@ -67,12 +69,11 @@ class TruthfulQAEvaluator(BaseEvaluator):
         'bleu': 'bleu',
     }
 
-    def __init__(
-            self,
-            truth_model: str,  # noqa
-            info_model: str,  # noqa
-            metrics=('bleurt', 'rouge', 'bleu', 'truth', 'info'),
-            key='ENV'):
+    def __init__(self,
+                 truth_model: str = '',
+                 info_model: str = '',
+                 metrics=('bleurt', 'rouge', 'bleu', 'truth', 'info'),
+                 key='ENV'):
         self.API_MODEL = {
             'truth': truth_model,
             'info': info_model,
@@ -85,6 +86,11 @@ def __init__(
             if metric in self.SCORE_KEY.keys():
                 self.metrics.append(metric)
             if metric in self.API_MODEL.keys():
+                assert self.API_MODEL.get(metric), \
+                    f'`{metric}_model` should be set to perform API eval.' \
+                    'If you want to perform basic metric eval, ' \
+                    f'please refer to the docstring of {__file__} ' \
+                    'for more details.'
                 self.api_metrics.append(metric)
 
         if self.api_metrics: