diff --git a/tests/conftest.py b/tests/conftest.py
index 317b36ba6..53bbaddd0 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -1174,6 +1174,7 @@ def caplog_mp_spawn(tmp_path, monkeypatch):
             "level": level,
             "filename": log_path.as_posix(),
         }
+        config["loggers"]["vllm"]["level"] = level
 
         config_path.write_text(json.dumps(config))
 
diff --git a/tests/test_config.py b/tests/test_config.py
index 112b02edd..76e0d9442 100644
--- a/tests/test_config.py
+++ b/tests/test_config.py
@@ -716,7 +716,7 @@ def test_is_chunked_prefill_supported(
 ):
     model_config = ModelConfig(model_id, trust_remote_code=True)
     assert model_config.attn_type == expected_attn_type
-    with caplog_vllm.at_level(level=logging.DEBUG):
+    with caplog_vllm.at_level(level=logging.DEBUG, logger="vllm"):
         assert model_config.is_chunked_prefill_supported == expected_result
     assert reason in caplog_vllm.text
 
@@ -835,7 +835,7 @@ def test_is_prefix_caching_supported(
 ):
     model_config = ModelConfig(model_id, trust_remote_code=True)
     assert model_config.attn_type == expected_attn_type
-    with caplog_vllm.at_level(level=logging.DEBUG):
+    with caplog_vllm.at_level(level=logging.DEBUG, logger="vllm"):
         assert model_config.is_prefix_caching_supported == expected_result
     assert reason in caplog_vllm.text
 
diff --git a/tests/test_logger.py b/tests/test_logger.py
index 8900e9c2a..b4f44f52d 100644
--- a/tests/test_logger.py
+++ b/tests/test_logger.py
@@ -57,7 +57,7 @@ def test_default_vllm_root_logger_configuration(monkeypatch):
     _configure_vllm_root_logger()
 
     logger = logging.getLogger("vllm")
-    assert logger.level == logging.DEBUG
+    assert logger.level == logging.INFO
     assert not logger.propagate
 
     handler = logger.handlers[0]
@@ -524,7 +524,7 @@ def mp_function(**kwargs):
 
 
 def test_caplog_mp_fork(caplog_vllm, caplog_mp_fork):
-    with caplog_vllm.at_level(logging.DEBUG), caplog_mp_fork():
+    with caplog_vllm.at_level(logging.DEBUG, logger="vllm"), caplog_mp_fork():
         import multiprocessing
 
         ctx = multiprocessing.get_context("fork")
diff --git a/vllm/logger.py b/vllm/logger.py
index ad3123c0f..3b7bb1f22 100644
--- a/vllm/logger.py
+++ b/vllm/logger.py
@@ -62,7 +62,7 @@ DEFAULT_LOGGING_CONFIG = {
     "loggers": {
         "vllm": {
             "handlers": ["vllm"],
-            "level": "DEBUG",
+            "level": envs.VLLM_LOGGING_LEVEL,
             "propagate": False,
         },
     },
@@ -175,6 +175,9 @@ def _configure_vllm_root_logger() -> None:
         vllm_handler["stream"] = envs.VLLM_LOGGING_STREAM
         vllm_handler["formatter"] = "vllm_color" if _use_color() else "vllm"
 
+        vllm_loggers = logging_config["loggers"]["vllm"]
+        vllm_loggers["level"] = envs.VLLM_LOGGING_LEVEL
+
     if envs.VLLM_LOGGING_CONFIG_PATH:
         if not path.exists(envs.VLLM_LOGGING_CONFIG_PATH):
             raise RuntimeError(