Add support for calling tools in Open Router (#148881)

homeassistant/components/open_router/config_flow.py

@@ -16,8 +16,9 @@ from homeassistant.config_entries import (
     ConfigSubentryFlow,
     SubentryFlowResult,
 )
-from homeassistant.const import CONF_API_KEY, CONF_MODEL
+from homeassistant.const import CONF_API_KEY, CONF_LLM_HASS_API, CONF_MODEL
 from homeassistant.core import callback
+from homeassistant.helpers import llm
 from homeassistant.helpers.aiohttp_client import async_get_clientsession
 from homeassistant.helpers.httpx_client import get_async_client
 from homeassistant.helpers.selector import (
@@ -25,9 +26,10 @@ from homeassistant.helpers.selector import (
     SelectSelector,
     SelectSelectorConfig,
     SelectSelectorMode,
+    TemplateSelector,
 )
 
-from .const import DOMAIN
+from .const import CONF_PROMPT, DOMAIN, RECOMMENDED_CONVERSATION_OPTIONS
 
 _LOGGER = logging.getLogger(__name__)
 
@@ -90,6 +92,8 @@ class ConversationFlowHandler(ConfigSubentryFlow):
     ) -> SubentryFlowResult:
         """User flow to create a sensor subentry."""
         if user_input is not None:
+            if not user_input.get(CONF_LLM_HASS_API):
+                user_input.pop(CONF_LLM_HASS_API, None)
             return self.async_create_entry(
                 title=self.options[user_input[CONF_MODEL]], data=user_input
             )
@@ -99,11 +103,17 @@ class ConversationFlowHandler(ConfigSubentryFlow):
             api_key=entry.data[CONF_API_KEY],
             http_client=get_async_client(self.hass),
         )
+        hass_apis: list[SelectOptionDict] = [
+            SelectOptionDict(
+                label=api.name,
+                value=api.id,
+            )
+            for api in llm.async_get_apis(self.hass)
+        ]
         options = []
         async for model in client.with_options(timeout=10.0).models.list():
             options.append(SelectOptionDict(value=model.id, label=model.name))  # type: ignore[attr-defined]
             self.options[model.id] = model.name  # type: ignore[attr-defined]
-
         return self.async_show_form(
             step_id="user",
             data_schema=vol.Schema(
@@ -113,6 +123,20 @@ class ConversationFlowHandler(ConfigSubentryFlow):
                             options=options, mode=SelectSelectorMode.DROPDOWN, sort=True
                         ),
                     ),
+                    vol.Optional(
+                        CONF_PROMPT,
+                        description={
+                            "suggested_value": RECOMMENDED_CONVERSATION_OPTIONS[
+                                CONF_PROMPT
+                            ]
+                        },
+                    ): TemplateSelector(),
+                    vol.Optional(
+                        CONF_LLM_HASS_API,
+                        default=RECOMMENDED_CONVERSATION_OPTIONS[CONF_LLM_HASS_API],
+                    ): SelectSelector(
+                        SelectSelectorConfig(options=hass_apis, multiple=True)
+                    ),
                 }
             ),
         )

homeassistant/components/open_router/const.py

@@ -2,5 +2,17 @@
 
 import logging
 
+from homeassistant.const import CONF_LLM_HASS_API
+from homeassistant.helpers import llm
+
 DOMAIN = "open_router"
 LOGGER = logging.getLogger(__package__)
+
+CONF_PROMPT = "prompt"
+CONF_RECOMMENDED = "recommended"
+
+RECOMMENDED_CONVERSATION_OPTIONS = {
+    CONF_RECOMMENDED: True,
+    CONF_LLM_HASS_API: [llm.LLM_API_ASSIST],
+    CONF_PROMPT: llm.DEFAULT_INSTRUCTIONS_PROMPT,
+}

homeassistant/components/open_router/conversation.py

@@ -1,25 +1,39 @@
 """Conversation support for OpenRouter."""
 
-from typing import Literal
+from collections.abc import AsyncGenerator, Callable
+import json
+from typing import Any, Literal
 
 import openai
+from openai import NOT_GIVEN
 from openai.types.chat import (
     ChatCompletionAssistantMessageParam,
+    ChatCompletionMessage,
     ChatCompletionMessageParam,
+    ChatCompletionMessageToolCallParam,
     ChatCompletionSystemMessageParam,
+    ChatCompletionToolMessageParam,
+    ChatCompletionToolParam,
     ChatCompletionUserMessageParam,
 )
+from openai.types.chat.chat_completion_message_tool_call_param import Function
+from openai.types.shared_params import FunctionDefinition
+from voluptuous_openapi import convert
 
 from homeassistant.components import conversation
 from homeassistant.config_entries import ConfigSubentry
 from homeassistant.const import CONF_LLM_HASS_API, CONF_MODEL, MATCH_ALL
 from homeassistant.core import HomeAssistant
 from homeassistant.exceptions import HomeAssistantError
+from homeassistant.helpers import llm
 from homeassistant.helpers.device_registry import DeviceEntryType, DeviceInfo
 from homeassistant.helpers.entity_platform import AddConfigEntryEntitiesCallback
 
 from . import OpenRouterConfigEntry
-from .const import DOMAIN, LOGGER
+from .const import CONF_PROMPT, DOMAIN, LOGGER
+
+# Max number of back and forth with the LLM to generate a response
+MAX_TOOL_ITERATIONS = 10
 
 
 async def async_setup_entry(
@@ -35,13 +49,31 @@ async def async_setup_entry(
         )
 
 
+def _format_tool(
+    tool: llm.Tool,
+    custom_serializer: Callable[[Any], Any] | None,
+) -> ChatCompletionToolParam:
+    """Format tool specification."""
+    tool_spec = FunctionDefinition(
+        name=tool.name,
+        parameters=convert(tool.parameters, custom_serializer=custom_serializer),
+    )
+    if tool.description:
+        tool_spec["description"] = tool.description
+    return ChatCompletionToolParam(type="function", function=tool_spec)
+
+
 def _convert_content_to_chat_message(
     content: conversation.Content,
 ) -> ChatCompletionMessageParam | None:
     """Convert any native chat message for this agent to the native format."""
     LOGGER.debug("_convert_content_to_chat_message=%s", content)
     if isinstance(content, conversation.ToolResultContent):
-        return None
+        return ChatCompletionToolMessageParam(
+            role="tool",
+            tool_call_id=content.tool_call_id,
+            content=json.dumps(content.tool_result),
+        )
 
     role: Literal["user", "assistant", "system"] = content.role
     if role == "system" and content.content:
@@ -51,13 +83,55 @@ def _convert_content_to_chat_message(
         return ChatCompletionUserMessageParam(role="user", content=content.content)
 
     if role == "assistant":
-        return ChatCompletionAssistantMessageParam(
+        param = ChatCompletionAssistantMessageParam(
-            role="assistant", content=content.content
+            role="assistant",
+            content=content.content,
         )
+        if isinstance(content, conversation.AssistantContent) and content.tool_calls:
+            param["tool_calls"] = [
+                ChatCompletionMessageToolCallParam(
+                    type="function",
+                    id=tool_call.id,
+                    function=Function(
+                        arguments=json.dumps(tool_call.tool_args),
+                        name=tool_call.tool_name,
+                    ),
+                )
+                for tool_call in content.tool_calls
+            ]
+        return param
     LOGGER.warning("Could not convert message to Completions API: %s", content)
     return None
 
 
+def _decode_tool_arguments(arguments: str) -> Any:
+    """Decode tool call arguments."""
+    try:
+        return json.loads(arguments)
+    except json.JSONDecodeError as err:
+        raise HomeAssistantError(f"Unexpected tool argument response: {err}") from err
+
+
+async def _transform_response(
+    message: ChatCompletionMessage,
+) -> AsyncGenerator[conversation.AssistantContentDeltaDict]:
+    """Transform the OpenRouter message to a ChatLog format."""
+    data: conversation.AssistantContentDeltaDict = {
+        "role": message.role,
+        "content": message.content,
+    }
+    if message.tool_calls:
+        data["tool_calls"] = [
+            llm.ToolInput(
+                id=tool_call.id,
+                tool_name=tool_call.function.name,
+                tool_args=_decode_tool_arguments(tool_call.function.arguments),
+            )
+            for tool_call in message.tool_calls
+        ]
+    yield data
+
+
 class OpenRouterConversationEntity(conversation.ConversationEntity):
     """OpenRouter conversation agent."""
 
@@ -75,6 +149,10 @@ class OpenRouterConversationEntity(conversation.ConversationEntity):
             name=subentry.title,
             entry_type=DeviceEntryType.SERVICE,
         )
+        if self.subentry.data.get(CONF_LLM_HASS_API):
+            self._attr_supported_features = (
+                conversation.ConversationEntityFeature.CONTROL
+            )
 
     @property
     def supported_languages(self) -> list[str] | Literal["*"]:
@@ -93,12 +171,19 @@ class OpenRouterConversationEntity(conversation.ConversationEntity):
             await chat_log.async_provide_llm_data(
                 user_input.as_llm_context(DOMAIN),
                 options.get(CONF_LLM_HASS_API),
-                None,
+                options.get(CONF_PROMPT),
                 user_input.extra_system_prompt,
             )
         except conversation.ConverseError as err:
             return err.as_conversation_result()
 
+        tools: list[ChatCompletionToolParam] | None = None
+        if chat_log.llm_api:
+            tools = [
+                _format_tool(tool, chat_log.llm_api.custom_serializer)
+                for tool in chat_log.llm_api.tools
+            ]
+
         messages = [
             m
             for content in chat_log.content
@@ -107,10 +192,12 @@ class OpenRouterConversationEntity(conversation.ConversationEntity):
 
         client = self.entry.runtime_data
 
+        for _iteration in range(MAX_TOOL_ITERATIONS):
             try:
                 result = await client.chat.completions.create(
                     model=self.model,
                     messages=messages,
+                    tools=tools or NOT_GIVEN,
                     user=chat_log.conversation_id,
                     extra_headers={
                         "X-Title": "Home Assistant",
@@ -123,11 +210,16 @@ class OpenRouterConversationEntity(conversation.ConversationEntity):
 
             result_message = result.choices[0].message
 
-        chat_log.async_add_assistant_content_without_tools(
+            messages.extend(
-            conversation.AssistantContent(
+                [
-                agent_id=user_input.agent_id,
+                    msg
-                content=result_message.content,
+                    async for content in chat_log.async_add_delta_content_stream(
+                        user_input.agent_id, _transform_response(result_message)
                     )
+                    if (msg := _convert_content_to_chat_message(content))
+                ]
             )
+            if not chat_log.unresponded_tool_results:
+                break
 
         return conversation.async_get_result_from_chat_log(user_input, chat_log)

homeassistant/components/open_router/strings.json

@@ -24,7 +24,13 @@
         "user": {
           "description": "Configure the new conversation agent",
           "data": {
-            "model": "Model"
+            "model": "Model",
+            "prompt": "Instructions",
+            "llm_hass_api": "[%key:common::config_flow::data::llm_hass_api%]"
+          },
+          "data_description": {
+            "model": "The model to use for the conversation agent",
+            "prompt": "Instruct how the LLM should respond. This can be a template."
           }
         }
       },

tests/components/open_router/conftest.py

@@ -2,6 +2,7 @@
 
 from collections.abc import AsyncGenerator, Generator
 from dataclasses import dataclass
+from typing import Any
 from unittest.mock import AsyncMock, MagicMock, patch
 
 from openai.types import CompletionUsage
@@ -9,10 +10,11 @@ from openai.types.chat import ChatCompletion, ChatCompletionMessage
 from openai.types.chat.chat_completion import Choice
 import pytest
 
-from homeassistant.components.open_router.const import DOMAIN
+from homeassistant.components.open_router.const import CONF_PROMPT, DOMAIN
 from homeassistant.config_entries import ConfigSubentryData
-from homeassistant.const import CONF_API_KEY, CONF_MODEL
+from homeassistant.const import CONF_API_KEY, CONF_LLM_HASS_API, CONF_MODEL
 from homeassistant.core import HomeAssistant
+from homeassistant.helpers import llm
 from homeassistant.setup import async_setup_component
 
 from tests.common import MockConfigEntry
@@ -29,7 +31,27 @@ def mock_setup_entry() -> Generator[AsyncMock]:
 
 
 @pytest.fixture
-def mock_config_entry(hass: HomeAssistant) -> MockConfigEntry:
+def enable_assist() -> bool:
+    """Mock conversation subentry data."""
+    return False
+
+
+@pytest.fixture
+def conversation_subentry_data(enable_assist: bool) -> dict[str, Any]:
+    """Mock conversation subentry data."""
+    res: dict[str, Any] = {
+        CONF_MODEL: "gpt-3.5-turbo",
+        CONF_PROMPT: "You are a helpful assistant.",
+    }
+    if enable_assist:
+        res[CONF_LLM_HASS_API] = [llm.LLM_API_ASSIST]
+    return res
+
+
+@pytest.fixture
+def mock_config_entry(
+    hass: HomeAssistant, conversation_subentry_data: dict[str, Any]
+) -> MockConfigEntry:
     """Mock a config entry."""
     return MockConfigEntry(
         title="OpenRouter",
@@ -39,7 +61,7 @@ def mock_config_entry(hass: HomeAssistant) -> MockConfigEntry:
         },
         subentries_data=[
             ConfigSubentryData(
-                data={CONF_MODEL: "gpt-3.5-turbo"},
+                data=conversation_subentry_data,
                 subentry_id="ABCDEF",
                 subentry_type="conversation",
                 title="GPT-3.5 Turbo",

tests/components/open_router/snapshots/test_conversation.ambr

@@ -1,4 +1,108 @@
 # serializer version: 1
+# name: test_all_entities[assist][conversation.gpt_3_5_turbo-entry]
+  EntityRegistryEntrySnapshot({
+    'aliases': set({
+    }),
+    'area_id': None,
+    'capabilities': None,
+    'config_entry_id': <ANY>,
+    'config_subentry_id': <ANY>,
+    'device_class': None,
+    'device_id': <ANY>,
+    'disabled_by': None,
+    'domain': 'conversation',
+    'entity_category': None,
+    'entity_id': 'conversation.gpt_3_5_turbo',
+    'has_entity_name': True,
+    'hidden_by': None,
+    'icon': None,
+    'id': <ANY>,
+    'labels': set({
+    }),
+    'name': None,
+    'options': dict({
+      'conversation': dict({
+        'should_expose': False,
+      }),
+    }),
+    'original_device_class': None,
+    'original_icon': None,
+    'original_name': None,
+    'platform': 'open_router',
+    'previous_unique_id': None,
+    'suggested_object_id': None,
+    'supported_features': <ConversationEntityFeature: 1>,
+    'translation_key': None,
+    'unique_id': 'ABCDEF',
+    'unit_of_measurement': None,
+  })
+# ---
+# name: test_all_entities[assist][conversation.gpt_3_5_turbo-state]
+  StateSnapshot({
+    'attributes': ReadOnlyDict({
+      'friendly_name': 'GPT-3.5 Turbo',
+      'supported_features': <ConversationEntityFeature: 1>,
+    }),
+    'context': <ANY>,
+    'entity_id': 'conversation.gpt_3_5_turbo',
+    'last_changed': <ANY>,
+    'last_reported': <ANY>,
+    'last_updated': <ANY>,
+    'state': 'unknown',
+  })
+# ---
+# name: test_all_entities[no_assist][conversation.gpt_3_5_turbo-entry]
+  EntityRegistryEntrySnapshot({
+    'aliases': set({
+    }),
+    'area_id': None,
+    'capabilities': None,
+    'config_entry_id': <ANY>,
+    'config_subentry_id': <ANY>,
+    'device_class': None,
+    'device_id': <ANY>,
+    'disabled_by': None,
+    'domain': 'conversation',
+    'entity_category': None,
+    'entity_id': 'conversation.gpt_3_5_turbo',
+    'has_entity_name': True,
+    'hidden_by': None,
+    'icon': None,
+    'id': <ANY>,
+    'labels': set({
+    }),
+    'name': None,
+    'options': dict({
+      'conversation': dict({
+        'should_expose': False,
+      }),
+    }),
+    'original_device_class': None,
+    'original_icon': None,
+    'original_name': None,
+    'platform': 'open_router',
+    'previous_unique_id': None,
+    'suggested_object_id': None,
+    'supported_features': 0,
+    'translation_key': None,
+    'unique_id': 'ABCDEF',
+    'unit_of_measurement': None,
+  })
+# ---
+# name: test_all_entities[no_assist][conversation.gpt_3_5_turbo-state]
+  StateSnapshot({
+    'attributes': ReadOnlyDict({
+      'friendly_name': 'GPT-3.5 Turbo',
+      'supported_features': <ConversationEntityFeature: 0>,
+    }),
+    'context': <ANY>,
+    'entity_id': 'conversation.gpt_3_5_turbo',
+    'last_changed': <ANY>,
+    'last_reported': <ANY>,
+    'last_updated': <ANY>,
+    'state': 'unknown',
+  })
+# ---
 # name: test_default_prompt
   list([
     dict({
@@ -14,3 +118,39 @@
     }),
   ])
 # ---
+# name: test_function_call[True]
+  list([
+    dict({
+      'attachments': None,
+      'content': 'Please call the test function',
+      'role': 'user',
+    }),
+    dict({
+      'agent_id': 'conversation.gpt_3_5_turbo',
+      'content': None,
+      'role': 'assistant',
+      'tool_calls': list([
+        dict({
+          'id': 'call_call_1',
+          'tool_args': dict({
+            'param1': 'call1',
+          }),
+          'tool_name': 'test_tool',
+        }),
+      ]),
+    }),
+    dict({
+      'agent_id': 'conversation.gpt_3_5_turbo',
+      'role': 'tool_result',
+      'tool_call_id': 'call_call_1',
+      'tool_name': 'test_tool',
+      'tool_result': 'value1',
+    }),
+    dict({
+      'agent_id': 'conversation.gpt_3_5_turbo',
+      'content': 'I have successfully called the function',
+      'role': 'assistant',
+      'tool_calls': None,
+    }),
+  ])
+# ---

tests/components/open_router/test_config_flow.py

@@ -5,9 +5,9 @@ from unittest.mock import AsyncMock
 import pytest
 from python_open_router import OpenRouterError
 
-from homeassistant.components.open_router.const import DOMAIN
+from homeassistant.components.open_router.const import CONF_PROMPT, DOMAIN
-from homeassistant.config_entries import SOURCE_USER, ConfigSubentry
+from homeassistant.config_entries import SOURCE_USER
-from homeassistant.const import CONF_API_KEY, CONF_MODEL
+from homeassistant.const import CONF_API_KEY, CONF_LLM_HASS_API, CONF_MODEL
 from homeassistant.core import HomeAssistant
 from homeassistant.data_entry_flow import FlowResultType
 
@@ -129,18 +129,56 @@ async def test_create_conversation_agent(
 
     result = await hass.config_entries.subentries.async_configure(
         result["flow_id"],
-        {CONF_MODEL: "gpt-3.5-turbo"},
+        {
+            CONF_MODEL: "gpt-3.5-turbo",
+            CONF_PROMPT: "you are an assistant",
+            CONF_LLM_HASS_API: ["assist"],
+        },
     )
 
     assert result["type"] is FlowResultType.CREATE_ENTRY
-    subentry_id = list(mock_config_entry.subentries)[0]
+    assert result["data"] == {
-    assert (
+        CONF_MODEL: "gpt-3.5-turbo",
-        ConfigSubentry(
+        CONF_PROMPT: "you are an assistant",
-            data={CONF_MODEL: "gpt-3.5-turbo"},
+        CONF_LLM_HASS_API: ["assist"],
-            subentry_id=subentry_id,
+    }
-            subentry_type="conversation",
+
-            title="GPT-3.5 Turbo",
+
-            unique_id=None,
+async def test_create_conversation_agent_no_control(
+    hass: HomeAssistant,
+    mock_open_router_client: AsyncMock,
+    mock_openai_client: AsyncMock,
+    mock_config_entry: MockConfigEntry,
+) -> None:
+    """Test creating a conversation agent without control over the LLM API."""
+
+    mock_config_entry.add_to_hass(hass)
+
+    await setup_integration(hass, mock_config_entry)
+
+    result = await hass.config_entries.subentries.async_init(
+        (mock_config_entry.entry_id, "conversation"),
+        context={"source": SOURCE_USER},
     )
-        in mock_config_entry.subentries.values()
+    assert result["type"] is FlowResultType.FORM
+    assert not result["errors"]
+    assert result["step_id"] == "user"
+
+    assert result["data_schema"].schema["model"].config["options"] == [
+        {"value": "gpt-3.5-turbo", "label": "GPT-3.5 Turbo"},
+    ]
+
+    result = await hass.config_entries.subentries.async_configure(
+        result["flow_id"],
+        {
+            CONF_MODEL: "gpt-3.5-turbo",
+            CONF_PROMPT: "you are an assistant",
+            CONF_LLM_HASS_API: [],
+        },
     )
+
+    assert result["type"] is FlowResultType.CREATE_ENTRY
+    assert result["data"] == {
+        CONF_MODEL: "gpt-3.5-turbo",
+        CONF_PROMPT: "you are an assistant",
+    }

tests/components/open_router/test_conversation.py

@@ -3,16 +3,24 @@
 from unittest.mock import AsyncMock
 
 from freezegun import freeze_time
+from openai.types import CompletionUsage
+from openai.types.chat import (
+    ChatCompletion,
+    ChatCompletionMessage,
+    ChatCompletionMessageToolCall,
+)
+from openai.types.chat.chat_completion import Choice
+from openai.types.chat.chat_completion_message_tool_call import Function
 import pytest
 from syrupy.assertion import SnapshotAssertion
 
 from homeassistant.components import conversation
 from homeassistant.core import Context, HomeAssistant
-from homeassistant.helpers import area_registry as ar, device_registry as dr, intent
+from homeassistant.helpers import entity_registry as er, intent
 
 from . import setup_integration
 
-from tests.common import MockConfigEntry
+from tests.common import MockConfigEntry, snapshot_platform
 from tests.components.conversation import MockChatLog, mock_chat_log  # noqa: F401
 
 
@@ -23,11 +31,23 @@ def freeze_the_time():
         yield
 
 
+@pytest.mark.parametrize("enable_assist", [True, False], ids=["assist", "no_assist"])
+async def test_all_entities(
+    hass: HomeAssistant,
+    snapshot: SnapshotAssertion,
+    mock_openai_client: AsyncMock,
+    mock_config_entry: MockConfigEntry,
+    entity_registry: er.EntityRegistry,
+) -> None:
+    """Test all entities."""
+    await setup_integration(hass, mock_config_entry)
+
+    await snapshot_platform(hass, entity_registry, snapshot, mock_config_entry.entry_id)
+
+
 async def test_default_prompt(
     hass: HomeAssistant,
     mock_config_entry: MockConfigEntry,
-    area_registry: ar.AreaRegistry,
-    device_registry: dr.DeviceRegistry,
     snapshot: SnapshotAssertion,
     mock_openai_client: AsyncMock,
     mock_chat_log: MockChatLog,  # noqa: F811
@@ -50,3 +70,95 @@ async def test_default_prompt(
         "HTTP-Referer": "https://www.home-assistant.io/integrations/open_router",
         "X-Title": "Home Assistant",
     }
+
+
+@pytest.mark.parametrize("enable_assist", [True])
+async def test_function_call(
+    hass: HomeAssistant,
+    mock_chat_log: MockChatLog,  # noqa: F811
+    mock_config_entry: MockConfigEntry,
+    snapshot: SnapshotAssertion,
+    mock_openai_client: AsyncMock,
+) -> None:
+    """Test function call from the assistant."""
+    await setup_integration(hass, mock_config_entry)
+
+    mock_chat_log.mock_tool_results(
+        {
+            "call_call_1": "value1",
+            "call_call_2": "value2",
+        }
+    )
+
+    async def completion_result(*args, messages, **kwargs):
+        for message in messages:
+            role = message["role"] if isinstance(message, dict) else message.role
+            if role == "tool":
+                return ChatCompletion(
+                    id="chatcmpl-1234567890ZYXWVUTSRQPONMLKJIH",
+                    choices=[
+                        Choice(
+                            finish_reason="stop",
+                            index=0,
+                            message=ChatCompletionMessage(
+                                content="I have successfully called the function",
+                                role="assistant",
+                                function_call=None,
+                                tool_calls=None,
+                            ),
+                        )
+                    ],
+                    created=1700000000,
+                    model="gpt-4-1106-preview",
+                    object="chat.completion",
+                    system_fingerprint=None,
+                    usage=CompletionUsage(
+                        completion_tokens=9, prompt_tokens=8, total_tokens=17
+                    ),
+                )
+
+        return ChatCompletion(
+            id="chatcmpl-1234567890ABCDEFGHIJKLMNOPQRS",
+            choices=[
+                Choice(
+                    finish_reason="tool_calls",
+                    index=0,
+                    message=ChatCompletionMessage(
+                        content=None,
+                        role="assistant",
+                        function_call=None,
+                        tool_calls=[
+                            ChatCompletionMessageToolCall(
+                                id="call_call_1",
+                                function=Function(
+                                    arguments='{"param1":"call1"}',
+                                    name="test_tool",
+                                ),
+                                type="function",
+                            )
+                        ],
+                    ),
+                )
+            ],
+            created=1700000000,
+            model="gpt-4-1106-preview",
+            object="chat.completion",
+            system_fingerprint=None,
+            usage=CompletionUsage(
+                completion_tokens=9, prompt_tokens=8, total_tokens=17
+            ),
+        )
+
+    mock_openai_client.chat.completions.create = completion_result
+
+    result = await conversation.async_converse(
+        hass,
+        "Please call the test function",
+        mock_chat_log.conversation_id,
+        Context(),
+        agent_id="conversation.gpt_3_5_turbo",
+    )
+
+    assert result.response.response_type == intent.IntentResponseType.ACTION_DONE
+    # Don't test the prompt, as it's not deterministic
+    assert mock_chat_log.content[1:] == snapshot