diff --git a/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n4g-dtensor2tp1.v1.sh b/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n4g-dtensor2tp1.v1.sh
index 842bfc9fa5..e4eeb2ded8 100755
--- a/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n4g-dtensor2tp1.v1.sh
+++ b/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n4g-dtensor2tp1.v1.sh
@@ -35,7 +35,7 @@ uv run tests/json_dump_tb_logs.py $LOG_DIR --output_path $JSON_METRICS
 # Only run metrics if the target step is reached
 if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | map(tonumber) | max' $JSON_METRICS) -ge $MAX_STEPS ]]; then
     uv run tests/check_metrics.py $JSON_METRICS \
-        'data["train/reward"]["200"] > 0.9'
+        'mean(data["train/reward"], range_start=181) > 0.9'
 
     # Clean up checkpoint directory after successful run to save space.
     rm -rf "$CKPT_DIR"
diff --git a/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n4g-megatrontp1.v1.sh b/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n4g-megatrontp1.v1.sh
index 7b15555457..2b915518ee 100755
--- a/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n4g-megatrontp1.v1.sh
+++ b/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n4g-megatrontp1.v1.sh
@@ -35,8 +35,8 @@ uv run tests/json_dump_tb_logs.py $LOG_DIR --output_path $JSON_METRICS
 # Only run metrics if the target step is reached
 if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | map(tonumber) | max' $JSON_METRICS) -ge $MAX_STEPS ]]; then
     uv run tests/check_metrics.py $JSON_METRICS \
-        'data["train/loss"]["200"] < 0.1' \
-        'data["train/reward"]["200"] > 0.9'
+        'mean(data["train/loss"], range_start=181) < 0.1' \
+        'mean(data["train/reward"], range_start=181) > 0.9'
 
     # Clean up checkpoint directory after successful run to save space.
     rm -rf "$CKPT_DIR"
diff --git a/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n8g-dtensor2tp1.v1.sh b/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n8g-dtensor2tp1.v1.sh
index 386263ba3d..b1b50b3c46 100755
--- a/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n8g-dtensor2tp1.v1.sh
+++ b/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n8g-dtensor2tp1.v1.sh
@@ -34,7 +34,7 @@ uv run tests/json_dump_tb_logs.py $LOG_DIR --output_path $JSON_METRICS
 # Only run metrics if the target step is reached
 if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | map(tonumber) | max' $JSON_METRICS) -ge $MAX_STEPS ]]; then
     uv run tests/check_metrics.py $JSON_METRICS \
-        'data["train/reward"]["200"] > 0.9'
+        'mean(data["train/reward"], range_start=181) > 0.9'
 
     # Clean up checkpoint directory after successful run to save space.
     rm -rf "$CKPT_DIR"
diff --git a/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n8g-megatrontp2.v1.sh b/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n8g-megatrontp2.v1.sh
index bbd490e4c0..3c42ce5407 100755
--- a/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n8g-megatrontp2.v1.sh
+++ b/tests/test_suites/vlm/vlm_grpo-qwen2.5-vl-3b-instruct-clevr-1n8g-megatrontp2.v1.sh
@@ -34,8 +34,8 @@ uv run tests/json_dump_tb_logs.py $LOG_DIR --output_path $JSON_METRICS
 # Only run metrics if the target step is reached
 if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | map(tonumber) | max' $JSON_METRICS) -ge $MAX_STEPS ]]; then
     uv run tests/check_metrics.py $JSON_METRICS \
-        'data["train/loss"]["200"] < 0.1' \
-        'data["train/reward"]["200"] > 0.9'
+        'mean(data["train/loss"], range_start=181) < 0.1' \
+        'mean(data["train/reward"], range_start=181) > 0.9'
 
     # Clean up checkpoint directory after successful run to save space.
     rm -rf "$CKPT_DIR"