diff --git a/src/openai/_utils/_logs.py b/src/openai/_utils/_logs.py
index c2c92e62f4..baf81d228b 100644
--- a/src/openai/_utils/_logs.py
+++ b/src/openai/_utils/_logs.py
@@ -8,6 +8,13 @@
 
 
 SENSITIVE_HEADERS = {"api-key", "authorization", "x-amz-security-token"}
+_LOG_LEVELS = {
+    "debug": logging.DEBUG,
+    "info": logging.INFO,
+    "warning": logging.WARNING,
+    "error": logging.ERROR,
+    "critical": logging.CRITICAL,
+}
 
 
 def _basic_config() -> None:
@@ -29,12 +36,15 @@ def setup_logging() -> None:
     # Transport loggers may include complete URLs. Leave their configuration to
     # the application instead of enabling them with the SDK's logging switch.
     env = os.environ.get("OPENAI_LOG")
-    if env == "debug":
-        _basic_config()
-        logger.setLevel(logging.DEBUG)
-    elif env == "info":
-        _basic_config()
-        logger.setLevel(logging.INFO)
+    if env is None:
+        return
+
+    level = _LOG_LEVELS.get(env)
+    if level is None:
+        return
+
+    _basic_config()
+    logger.setLevel(level)
 
 
 class SensitiveHeadersFilter(logging.Filter):
@@ -45,4 +55,4 @@ def filter(self, record: logging.LogRecord) -> bool:
             for header in headers:
                 if str(header).lower() in SENSITIVE_HEADERS:
                     headers[header] = "<redacted>"
-        return True
+        return True
\ No newline at end of file
diff --git a/tests/test_log_levels.py b/tests/test_log_levels.py
new file mode 100644
index 0000000000..e36f7005e1
--- /dev/null
+++ b/tests/test_log_levels.py
@@ -0,0 +1,33 @@
+from __future__ import annotations
+
+import logging
+
+import pytest
+
+from openai._utils._logs import setup_logging
+
+
+@pytest.mark.parametrize(
+    ("setting", "expected_level"),
+    [
+        ("debug", logging.DEBUG),
+        ("info", logging.INFO),
+        ("warning", logging.WARNING),
+        ("error", logging.ERROR),
+        ("critical", logging.CRITICAL),
+    ],
+)
+def test_openai_log_sets_standard_log_level(
+    setting: str,
+    expected_level: int,
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    logger = logging.getLogger("openai")
+    original_level = logger.level
+
+    try:
+        monkeypatch.setenv("OPENAI_LOG", setting)
+        setup_logging()
+        assert logger.level == expected_level
+    finally:
+        logger.setLevel(original_level)