diff --git a/evaluation/README.md b/evaluation/README.md
index a5a4f32ca..8c1896943 100644
--- a/evaluation/README.md
+++ b/evaluation/README.md
@@ -64,6 +64,12 @@ First prepare the dataset `longmemeval_s` from https://huggingface.co/datasets/x
 ./scripts/run_lme_eval.sh
 ```
 
+#### Question date and `reference_time`
+
+LongMemEval gives each question a **question date**; evaluation should use that as the reference “now”, not the time when you run the script. The LongMemEval search script passes `question_date` as **`reference_time`** where the backend supports it.
+
+**MemOS Cloud** currently does not support supplying question date on search the same way, so LongMemEval scores there may differ from a spec-faithful run. **Prefer evaluating LongMemEval against the open-source MemOS server** when you need comparable numbers.
+
 ### PrefEval Evaluation
 Downloading benchmark_dataset/filtered_inter_turns.json from https://github.com/amazon-science/PrefEval/blob/main/benchmark_dataset/filtered_inter_turns.json and save it as `./data/prefeval/filtered_inter_turns.json`.
 To evaluate the **Prefeval** dataset — run the following [script](./scripts/run_prefeval_eval.sh):
diff --git a/evaluation/scripts/longmemeval/lme_search.py b/evaluation/scripts/longmemeval/lme_search.py
index 8e0e3c5c2..1eea8cd37 100644
--- a/evaluation/scripts/longmemeval/lme_search.py
+++ b/evaluation/scripts/longmemeval/lme_search.py
@@ -41,9 +41,11 @@ def mem0_search(client, query, user_id, top_k):
     return context, duration_ms
 
 
-def memos_search(client, query, user_id, top_k):
+def memos_search(client, query, user_id, top_k, reference_time=None):
     start = time()
-    results = client.search(query=query, user_id=user_id, top_k=top_k)
+    results = client.search(
+        query=query, user_id=user_id, top_k=top_k, reference_time=reference_time
+    )
     context = (
         "\n".join([i["memory"] for i in results["text_mem"][0]["memories"]])
         + f"\n{results.get('pref_string', '')}"
@@ -122,12 +124,16 @@ def process_user(lme_df, conv_idx, frame, version, top_k=20):
         from utils.client import MemosApiClient
 
         client = MemosApiClient()
-        context, duration_ms = memos_search(client, question, user_id, top_k)
+        context, duration_ms = memos_search(
+            client, question, user_id, top_k, reference_time=question_date
+        )
     elif frame == "memos-api-online":
         from utils.client import MemosApiOnlineClient
 
         client = MemosApiOnlineClient()
-        context, duration_ms = memos_search(client, question, user_id, top_k)
+        context, duration_ms = memos_search(
+            client, question, user_id, top_k, reference_time=question_date
+        )
     elif frame == "memu":
         from utils.client import MemuClient