diff --git a/README.md b/README.md
index f732655..557c75d 100644
--- a/README.md
+++ b/README.md
@@ -33,7 +33,7 @@ pip install fish-audio-sdk[utils]
 
 ## Authentication
 
-Get your API key from [fish.audio/app/api-keys](https://fish.audio/app/api-keys):
+Get your API key from [fish.audio/app/api-keys](https://fish.audio/app/api-keys/?utm_source=github&utm_medium=sdk-readme&utm_content=python):
 
 ```bash
 export FISH_API_KEY=your_api_key_here
@@ -93,6 +93,18 @@ production_audio = client.tts.convert(
     text="Production speech",
     model="s2.1-pro",
 )
+
+# Free to use, including on a new account with no API credit
+free_audio = client.tts.convert(
+    text="Free speech",
+    model="s2.1-pro-free",
+)
+
+# Drama 3: our most expressive model (uses API credit)
+drama_audio = client.tts.convert(
+    text="Expressive speech",
+    model="drama-3",
+)
 ```
 
 **With custom voice:**
diff --git a/src/fishaudio/types/shared.py b/src/fishaudio/types/shared.py
index a160e65..0628f9f 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["s1", "s2-pro", "s2.1-pro"]
+Model = Literal["s1", "s2-pro", "s2.1-pro", "s2.1-pro-free", "drama-3"]
 
 # Deprecated models
 DEPRECATED_MODELS = {"speech-1.5", "speech-1.6"}
diff --git a/src/fishaudio/types/tts.py b/src/fishaudio/types/tts.py
index a6bfbc3..8799cbd 100644
--- a/src/fishaudio/types/tts.py
+++ b/src/fishaudio/types/tts.py
@@ -2,7 +2,7 @@
 
 from typing import Annotated, Literal, Optional
 
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, Field, field_serializer
 
 from .shared import AudioFormat, LatencyMode
 
@@ -157,6 +157,11 @@ class TTSRequest(BaseModel):
     condition_on_previous_chunks: bool = True
     early_stop_threshold: float = 1.0
 
+    @field_serializer("opus_bitrate")
+    def _opus_bitrate_bps(self, value: int) -> int:
+        # The API takes opus_bitrate in bps (24000, 32000, ...); -1000 means auto.
+        return value * 1000 if value > 0 else value
+
 
 # WebSocket event types for streaming TTS
 class StartEvent(BaseModel):
diff --git a/tests/unit/test_tts.py b/tests/unit/test_tts.py
index f6075a5..7cabc48 100644
--- a/tests/unit/test_tts.py
+++ b/tests/unit/test_tts.py
@@ -87,6 +87,30 @@ def test_convert_basic(self, tts_client, mock_client_wrapper):
         call_args = mock_client_wrapper.request.call_args
         assert call_args[1]["headers"]["model"] == "s2.1-pro"
 
+    def test_convert_with_free_model(self, tts_client, mock_client_wrapper):
+        """Test the free s2.1-pro-free model is sent as the model header."""
+        mock_response = Mock()
+        mock_response.iter_bytes.return_value = iter([b"audio"])
+        mock_client_wrapper.request.return_value = mock_response
+
+        audio = tts_client.convert(text="Hello world", model="s2.1-pro-free")
+
+        assert audio == b"audio"
+        call_args = mock_client_wrapper.request.call_args
+        assert call_args[1]["headers"]["model"] == "s2.1-pro-free"
+
+    def test_convert_with_drama_3(self, tts_client, mock_client_wrapper):
+        """Test the drama-3 model is sent as the model header."""
+        mock_response = Mock()
+        mock_response.iter_bytes.return_value = iter([b"audio"])
+        mock_client_wrapper.request.return_value = mock_response
+
+        audio = tts_client.convert(text="Hello world", model="drama-3")
+
+        assert audio == b"audio"
+        call_args = mock_client_wrapper.request.call_args
+        assert call_args[1]["headers"]["model"] == "drama-3"
+
     def test_convert_with_reference_id(self, tts_client, mock_client_wrapper):
         """Test TTS with reference voice ID."""
         mock_response = Mock()
diff --git a/tests/unit/test_types.py b/tests/unit/test_types.py
index cfc8446..d6f760a 100644
--- a/tests/unit/test_types.py
+++ b/tests/unit/test_types.py
@@ -173,3 +173,11 @@ def test_voice_with_unlist_visibility(self, sample_voice_response):
         sample_voice_response["visibility"] = "unlist"
         voice = Voice.model_validate(sample_voice_response)
         assert voice.visibility == "unlist"
+
+
+def test_tts_request_sends_opus_bitrate_in_bps():
+    assert TTSRequest(text="hi", format="opus").model_dump()["opus_bitrate"] == 32000
+    assert TTSRequest(text="hi", opus_bitrate=64).model_dump()["opus_bitrate"] == 64000
+    assert (
+        TTSRequest(text="hi", opus_bitrate=-1000).model_dump()["opus_bitrate"] == -1000
+    )