diff --git a/README.md b/README.md
index c04d204..f732655 100644
--- a/README.md
+++ b/README.md
@@ -85,6 +85,16 @@ asyncio.run(main())
 
 ### Text-to-Speech
 
+**Selecting a model:**
+
+```python
+# Recommended for production
+production_audio = client.tts.convert(
+    text="Production speech",
+    model="s2.1-pro",
+)
+```
+
 **With custom voice:**
 
 ```python
diff --git a/src/fishaudio/resources/tts.py b/src/fishaudio/resources/tts.py
index 0167ce6..2a5114a 100644
--- a/src/fishaudio/resources/tts.py
+++ b/src/fishaudio/resources/tts.py
@@ -82,7 +82,7 @@ def stream(
         latency: Optional[LatencyMode] = None,
         speed: Optional[float] = None,
         config: TTSConfig = TTSConfig(),
-        model: Model = "s2-pro",
+        model: Union[Model, str] = "s2.1-pro",
         request_options: Optional[RequestOptions] = None,
     ) -> AudioStream:
         """
@@ -166,7 +166,7 @@ def convert(
         latency: Optional[LatencyMode] = None,
         speed: Optional[float] = None,
         config: TTSConfig = TTSConfig(),
-        model: Model = "s2-pro",
+        model: Union[Model, str] = "s2.1-pro",
         request_options: Optional[RequestOptions] = None,
     ) -> bytes:
         """
@@ -228,7 +228,7 @@ def stream_websocket(
         latency: Optional[LatencyMode] = None,
         speed: Optional[float] = None,
         config: TTSConfig = TTSConfig(),
-        model: Model = "s2-pro",
+        model: Union[Model, str] = "s2.1-pro",
         max_workers: int = 10,
         ws_options: Optional[WebSocketOptions] = None,
     ) -> Iterator[bytes]:
@@ -386,7 +386,7 @@ async def stream(
         latency: Optional[LatencyMode] = None,
         speed: Optional[float] = None,
         config: TTSConfig = TTSConfig(),
-        model: Model = "s2-pro",
+        model: Union[Model, str] = "s2.1-pro",
         request_options: Optional[RequestOptions] = None,
     ) -> AsyncAudioStream:
         """
@@ -471,7 +471,7 @@ async def convert(
         latency: Optional[LatencyMode] = None,
         speed: Optional[float] = None,
         config: TTSConfig = TTSConfig(),
-        model: Model = "s2-pro",
+        model: Union[Model, str] = "s2.1-pro",
         request_options: Optional[RequestOptions] = None,
     ) -> bytes:
         """
@@ -534,7 +534,7 @@ async def stream_websocket(
         latency: Optional[LatencyMode] = None,
         speed: Optional[float] = None,
         config: TTSConfig = TTSConfig(),
-        model: Model = "s2-pro",
+        model: Union[Model, str] = "s2.1-pro",
         ws_options: Optional[WebSocketOptions] = None,
     ):
         """
diff --git a/src/fishaudio/types/shared.py b/src/fishaudio/types/shared.py
index 879a012..a160e65 100644
--- a/src/fishaudio/types/shared.py
+++ b/src/fishaudio/types/shared.py
@@ -22,7 +22,7 @@ class PaginatedResponse(BaseModel, Generic[T]):
 
 
 # Model types
-Model = Literal["speech-1.5", "speech-1.6", "s1", "s2-pro"]
+Model = Literal["s1", "s2-pro", "s2.1-pro"]
 
 # Deprecated models
 DEPRECATED_MODELS = {"speech-1.5", "speech-1.6"}
@@ -32,7 +32,7 @@ def warn_if_deprecated_model(model: str) -> None:
     """Emit a deprecation warning if a legacy model is used."""
     if model in DEPRECATED_MODELS:
         warnings.warn(
-            f"Model '{model}' is deprecated. Use 's1' or 's2-pro' instead.",
+            f"Model '{model}' is deprecated. Use 's2.1-pro' instead.",
             DeprecationWarning,
             stacklevel=3,
         )
diff --git a/tests/unit/test_tts.py b/tests/unit/test_tts.py
index 132038a..f6075a5 100644
--- a/tests/unit/test_tts.py
+++ b/tests/unit/test_tts.py
@@ -63,7 +63,7 @@ def test_stream_basic(self, tts_client, mock_client_wrapper):
 
         # Check headers
         assert call_args[1]["headers"]["Content-Type"] == "application/msgpack"
-        assert call_args[1]["headers"]["model"] == "s2-pro"  # default model
+        assert call_args[1]["headers"]["model"] == "s2.1-pro"  # default model
 
         # Check payload was msgpack encoded
         assert "content" in call_args[1]
@@ -84,6 +84,8 @@ def test_convert_basic(self, tts_client, mock_client_wrapper):
 
         # Verify request was made correctly
         mock_client_wrapper.request.assert_called_once()
+        call_args = mock_client_wrapper.request.call_args
+        assert call_args[1]["headers"]["model"] == "s2.1-pro"
 
     def test_convert_with_reference_id(self, tts_client, mock_client_wrapper):
         """Test TTS with reference voice ID."""
@@ -510,6 +512,7 @@ async def async_iter_bytes():
 
         assert call_args[0][0] == "POST"
         assert call_args[0][1] == "/v1/tts"
+        assert call_args[1]["headers"]["model"] == "s2.1-pro"
 
     @pytest.mark.asyncio
     async def test_convert_basic(self, async_tts_client, async_mock_client_wrapper):
@@ -533,6 +536,8 @@ async def async_iter_bytes():
 
         # Verify request was made
         async_mock_client_wrapper.request.assert_called_once()
+        call_args = async_mock_client_wrapper.request.call_args
+        assert call_args[1]["headers"]["model"] == "s2.1-pro"
 
     @pytest.mark.asyncio
     async def test_convert_with_reference_id(
diff --git a/tests/unit/test_tts_realtime.py b/tests/unit/test_tts_realtime.py
index c3b60a9..0a25e62 100644
--- a/tests/unit/test_tts_realtime.py
+++ b/tests/unit/test_tts_realtime.py
@@ -92,6 +92,7 @@ def test_stream_websocket_basic(
             # Verify WebSocket connection was created
             mock_connect_ws.assert_called_once()
             assert mock_connect_ws.call_args[0][0] == "/v1/tts/live"
+            assert mock_connect_ws.call_args[1]["headers"]["model"] == "s2.1-pro"
 
     @pytest.mark.filterwarnings("ignore::DeprecationWarning")
     @patch("fishaudio.resources.tts.connect_ws")
@@ -425,6 +426,7 @@ async def text_stream():
             # Verify WebSocket connection was created
             mock_aconnect_ws.assert_called_once()
             assert mock_aconnect_ws.call_args[0][0] == "/v1/tts/live"
+            assert mock_aconnect_ws.call_args[1]["headers"]["model"] == "s2.1-pro"
 
     @pytest.mark.filterwarnings("ignore::DeprecationWarning")
     @pytest.mark.asyncio