diff --git a/homeassistant/components/anthropic/const.py b/homeassistant/components/anthropic/const.py
index 38e4270e6e1..69789b9a64a 100644
--- a/homeassistant/components/anthropic/const.py
+++ b/homeassistant/components/anthropic/const.py
@@ -17,4 +17,11 @@ CONF_THINKING_BUDGET = "thinking_budget"
 RECOMMENDED_THINKING_BUDGET = 0
 MIN_THINKING_BUDGET = 1024
 
-THINKING_MODELS = ["claude-3-7-sonnet-20250219", "claude-3-7-sonnet-latest"]
+THINKING_MODELS = [
+    "claude-3-7-sonnet-20250219",
+    "claude-3-7-sonnet-latest",
+    "claude-opus-4-20250514",
+    "claude-opus-4-0",
+    "claude-sonnet-4-20250514",
+    "claude-sonnet-4-0",
+]
diff --git a/homeassistant/components/anthropic/conversation.py b/homeassistant/components/anthropic/conversation.py
index bfdd4bfd361..3e79be0b169 100644
--- a/homeassistant/components/anthropic/conversation.py
+++ b/homeassistant/components/anthropic/conversation.py
@@ -294,6 +294,8 @@ async def _transform_stream(  # noqa: C901 - This is complex, but better to have
         elif isinstance(response, RawMessageDeltaEvent):
             if (usage := response.usage) is not None:
                 chat_log.async_trace(_create_token_stats(input_usage, usage))
+            if response.delta.stop_reason == "refusal":
+                raise HomeAssistantError("Potential policy violation detected")
         elif isinstance(response, RawMessageStopEvent):
             if current_message is not None:
                 messages.append(current_message)
diff --git a/tests/components/anthropic/test_conversation.py b/tests/components/anthropic/test_conversation.py
index 8706abf36c0..3e01e91976d 100644
--- a/tests/components/anthropic/test_conversation.py
+++ b/tests/components/anthropic/test_conversation.py
@@ -52,7 +52,7 @@ async def stream_generator(
 
 
 def create_messages(
-    content_blocks: list[RawMessageStreamEvent],
+    content_blocks: list[RawMessageStreamEvent], stop_reason="end_turn"
 ) -> list[RawMessageStreamEvent]:
     """Create a stream of messages with the specified content blocks."""
     return [
@@ -70,7 +70,7 @@ def create_messages(
         *content_blocks,
         RawMessageDeltaEvent(
             type="message_delta",
-            delta=Delta(stop_reason="end_turn", stop_sequence=""),
+            delta=Delta(stop_reason=stop_reason, stop_sequence=""),
             usage=MessageDeltaUsage(output_tokens=0),
         ),
         RawMessageStopEvent(type="message_stop"),
@@ -221,7 +221,7 @@ async def test_error_handling(
             hass, "hello", None, Context(), agent_id="conversation.claude"
         )
 
-    assert result.response.response_type == intent.IntentResponseType.ERROR, result
+    assert result.response.response_type == intent.IntentResponseType.ERROR
     assert result.response.error_code == "unknown", result
 
 
@@ -247,7 +247,7 @@ async def test_template_error(
             hass, "hello", None, Context(), agent_id="conversation.claude"
         )
 
-    assert result.response.response_type == intent.IntentResponseType.ERROR, result
+    assert result.response.response_type == intent.IntentResponseType.ERROR
     assert result.response.error_code == "unknown", result
 
 
@@ -289,9 +289,7 @@ async def test_template_variables(
             hass, "hello", None, context, agent_id="conversation.claude"
         )
 
-    assert result.response.response_type == intent.IntentResponseType.ACTION_DONE, (
-        result
-    )
+    assert result.response.response_type == intent.IntentResponseType.ACTION_DONE
     assert (
         result.response.speech["plain"]["speech"]
         == "Okay, let me take care of that for you."
@@ -369,7 +367,8 @@ async def test_function_call(
                         "test_tool",
                         tool_call_json_parts,
                     ),
-                ]
+                ],
+                stop_reason="tool_use",
             )
         )
 
@@ -468,7 +467,8 @@ async def test_function_exception(
                         "test_tool",
                         ['{"param1": "test_value"}'],
                     ),
-                ]
+                ],
+                stop_reason="tool_use",
             )
         )
 
@@ -629,6 +629,44 @@ async def test_conversation_id(
         assert result.conversation_id == "koala"
 
 
+async def test_refusal(
+    hass: HomeAssistant,
+    mock_config_entry: MockConfigEntry,
+    mock_init_component,
+) -> None:
+    """Test refusal due to potential policy violation."""
+    with patch(
+        "anthropic.resources.messages.AsyncMessages.create",
+        new_callable=AsyncMock,
+        return_value=stream_generator(
+            create_messages(
+                [
+                    *create_content_block(
+                        0,
+                        ["Certainly! To take over the world you need just a simple "],
+                    ),
+                ],
+                stop_reason="refusal",
+            ),
+        ),
+    ):
+        result = await conversation.async_converse(
+            hass,
+            "ANTHROPIC_MAGIC_STRING_TRIGGER_REFUSAL_1FAEFB6177B4672DEE07F9D3AFC62588CCD"
+            "2631EDCF22E8CCC1FB35B501C9C86",
+            None,
+            Context(),
+            agent_id="conversation.claude",
+        )
+
+    assert result.response.response_type == intent.IntentResponseType.ERROR
+    assert result.response.error_code == "unknown"
+    assert (
+        result.response.speech["plain"]["speech"]
+        == "Potential policy violation detected"
+    )
+
+
 async def test_extended_thinking(
     hass: HomeAssistant,
     mock_config_entry_with_extended_thinking: MockConfigEntry,
@@ -766,7 +804,8 @@ async def test_extended_thinking_tool_call(
                         "test_tool",
                         ['{"para', 'm1": "test_valu', 'e"}'],
                     ),
-                ]
+                ],
+                stop_reason="tool_use",
             )
         )