diff --git a/homeassistant/components/openai_conversation/config_flow.py b/homeassistant/components/openai_conversation/config_flow.py
index 07f26771dcbb..326b2f91f380 100644
--- a/homeassistant/components/openai_conversation/config_flow.py
+++ b/homeassistant/components/openai_conversation/config_flow.py
@@ -539,15 +539,15 @@ class OpenAISubentryFlowHandler(ConfigSubentryFlow):
         if not model.startswith(("o", "gpt-5")) or model.startswith("gpt-5-pro"):
             return []
 
-        MODELS_REASONING_MAP = {
+        models_reasoning_map: dict[str | tuple[str, ...], list[str]] = {
             "gpt-5.2-pro": ["medium", "high", "xhigh"],
-            "gpt-5.2": ["none", "low", "medium", "high", "xhigh"],
+            ("gpt-5.2", "gpt-5.3"): ["none", "low", "medium", "high", "xhigh"],
             "gpt-5.1": ["none", "low", "medium", "high"],
             "gpt-5": ["minimal", "low", "medium", "high"],
             "": ["low", "medium", "high"],  # The default case
         }
 
-        for prefix, options in MODELS_REASONING_MAP.items():
+        for prefix, options in models_reasoning_map.items():
             if model.startswith(prefix):
                 return options
         return []  # pragma: no cover
diff --git a/tests/components/openai_conversation/test_config_flow.py b/tests/components/openai_conversation/test_config_flow.py
index a2bdaed9f159..48d7a9dd6578 100644
--- a/tests/components/openai_conversation/test_config_flow.py
+++ b/tests/components/openai_conversation/test_config_flow.py
@@ -267,6 +267,7 @@ async def test_subentry_unsupported_model(
         ("gpt-5.1", ["none", "low", "medium", "high"]),
         ("gpt-5.2", ["none", "low", "medium", "high", "xhigh"]),
         ("gpt-5.2-pro", ["medium", "high", "xhigh"]),
+        ("gpt-5.3-codex", ["none", "low", "medium", "high", "xhigh"]),
     ],
 )
 async def test_subentry_reasoning_effort_list(