fix more tests

Author: Kludex

pydantic_ai_slim/pydantic_ai/models/function.py

@@ -135,6 +135,7 @@ async def request(
             assert isinstance(response_, ModelResponse), response_
             response = response_
         response.model_name = self._model_name
+        response.provider_name = self._system
         # Add usage data if not already present
         if not response.usage.has_values():  # pragma: no branch
             response.usage = _estimate_usage(chain(messages, [response]))
@@ -169,6 +170,7 @@ async def request_stream(
             model_request_parameters=model_request_parameters,
             _model_name=self._model_name,
             _iter=response_stream,
+            _provider_name=self._system,
         )
 
     @property
@@ -261,6 +263,7 @@ class FunctionStreamedResponse(StreamedResponse):
     _model_name: str
     _iter: AsyncIterator[str | DeltaToolCalls | DeltaThinkingCalls]
     _timestamp: datetime = field(default_factory=_utils.now_utc)
+    _provider_name: str
 
     def __post_init__(self):
         self._usage += _estimate_usage([])
@@ -305,9 +308,9 @@ def model_name(self) -> str:
         return self._model_name
 
     @property
-    def provider_name(self) -> None:
+    def provider_name(self) -> str:
         """Get the provider name."""
-        return None
+        return self._provider_name
 
     @property
     def timestamp(self) -> datetime:

pydantic_ai_slim/pydantic_ai/models/test.py

@@ -100,8 +100,6 @@ def __init__(
         self.custom_output_args = custom_output_args
         self.seed = seed
         self.last_model_request_parameters = None
-        self._model_name = 'test'
-        self._system = 'test'
         super().__init__(settings=settings, profile=profile)
 
     async def request(

tests/models/cassettes/test_model_names/test_known_model_names.yaml

@@ -109,39 +109,39 @@ interactions:
     parsed_body:
       data:
       - created: 0
-        id: llama-3.3-70b
+        id: qwen-3-235b-a22b-thinking-2507
         object: model
         owned_by: Cerebras
       - created: 0
-        id: qwen-3-235b-a22b-instruct-2507
+        id: qwen-3-coder-480b
         object: model
         owned_by: Cerebras
       - created: 0
-        id: gpt-oss-120b
+        id: llama3.1-8b
         object: model
         owned_by: Cerebras
       - created: 0
-        id: qwen-3-235b-a22b-thinking-2507
+        id: llama-4-scout-17b-16e-instruct
         object: model
         owned_by: Cerebras
       - created: 0
-        id: qwen-3-coder-480b
+        id: qwen-3-235b-a22b-instruct-2507
         object: model
         owned_by: Cerebras
       - created: 0
-        id: llama3.1-8b
+        id: qwen-3-32b
         object: model
         owned_by: Cerebras
       - created: 0
-        id: qwen-3-32b
+        id: llama-3.3-70b
         object: model
         owned_by: Cerebras
       - created: 0
-        id: llama-4-scout-17b-16e-instruct
+        id: llama-4-maverick-17b-128e-instruct
         object: model
         owned_by: Cerebras
       - created: 0
-        id: llama-4-maverick-17b-128e-instruct
+        id: gpt-oss-120b
         object: model
         owned_by: Cerebras
       object: list

tests/models/cassettes/test_openai/test_request_simple_usage.yaml

@@ -0,0 +1,80 @@
+interactions:
+- request:
+    headers:
+      accept:
+      - application/json
+      accept-encoding:
+      - gzip, deflate
+      connection:
+      - keep-alive
+      content-length:
+      - '100'
+      content-type:
+      - application/json
+      host:
+      - api.openai.com
+    method: POST
+    parsed_body:
+      messages:
+      - content: Hello! How are you doing?
+        role: user
+      model: gpt-4o
+      stream: false
+    uri: https://api.openai.com/v1/chat/completions
+  response:
+    headers:
+      access-control-expose-headers:
+      - X-Request-ID
+      alt-svc:
+      - h3=":443"; ma=86400
+      connection:
+      - keep-alive
+      content-length:
+      - '927'
+      content-type:
+      - application/json
+      openai-organization:
+      - pydantic-28gund
+      openai-processing-ms:
+      - '825'
+      openai-project:
+      - proj_dKobscVY9YJxeEaDJen54e3d
+      openai-version:
+      - '2020-10-01'
+      strict-transport-security:
+      - max-age=31536000; includeSubDomains; preload
+      transfer-encoding:
+      - chunked
+    parsed_body:
+      choices:
+      - finish_reason: stop
+        index: 0
+        logprobs: null
+        message:
+          annotations: []
+          content: Hello! I'm just a computer program, so I don't have feelings, but I'm here and ready to help you. How can
+            I assist you today?
+          refusal: null
+          role: assistant
+      created: 1756380852
+      id: chatcmpl-C9VBMiEq0GYAxsZn9U6FURAbS9Lau
+      model: gpt-4o-2024-08-06
+      object: chat.completion
+      service_tier: default
+      system_fingerprint: fp_07871e2ad8
+      usage:
+        completion_tokens: 30
+        completion_tokens_details:
+          accepted_prediction_tokens: 0
+          audio_tokens: 0
+          reasoning_tokens: 0
+          rejected_prediction_tokens: 0
+        prompt_tokens: 14
+        prompt_tokens_details:
+          audio_tokens: 0
+          cached_tokens: 0
+        total_tokens: 44
+    status:
+      code: 200
+      message: OK
+version: 1

tests/models/mock_openai.py

@@ -87,7 +87,7 @@ def completion_message(
         id='123',
         choices=choices,
         created=1704067200,  # 2024-01-01
-        model='gpt-4o-123',
+        model='gpt-4o',
         object='chat.completion',
         usage=usage,
     )
@@ -150,7 +150,7 @@ def response_message(
 ) -> responses.Response:
     return responses.Response(
         id='123',
-        model='gpt-4o-123',
+        model='gpt-4o',
         object='response',
         created_at=1704067200,  # 2024-01-01
         output=list(output_items),

tests/models/test_openai.py

@@ -87,6 +87,12 @@
     pytest.mark.skipif(not imports_successful(), reason='openai not installed'),
     pytest.mark.anyio,
     pytest.mark.vcr,
+    # TODO(Marcelo): genai-prices needs to include Cerebras prices: https://github.com/pydantic/genai-prices/issues/132
+    pytest.mark.filterwarnings('ignore:The costs with provider "cerebras" and model:UserWarning'),
+    # NOTE(Marcelo): The following model is old, so we are probably not including it on `genai-prices`.
+    pytest.mark.filterwarnings(
+        'ignore:The costs with provider "openai" and model "gpt-4o-search-preview-2025-03-11":UserWarning'
+    ),
 ]
 
 
@@ -120,15 +126,15 @@ async def test_request_simple_success(allow_model_requests: None):
             ModelRequest(parts=[UserPromptPart(content='hello', timestamp=IsNow(tz=timezone.utc))]),
             ModelResponse(
                 parts=[TextPart(content='world')],
-                model_name='gpt-4o-123',
+                model_name='gpt-4o',
                 timestamp=datetime(2024, 1, 1, 0, 0, tzinfo=timezone.utc),
                 provider_name='openai',
                 provider_request_id='123',
             ),
             ModelRequest(parts=[UserPromptPart(content='hello', timestamp=IsNow(tz=timezone.utc))]),
             ModelResponse(
                 parts=[TextPart(content='world')],
-                model_name='gpt-4o-123',
+                model_name='gpt-4o',
                 timestamp=datetime(2024, 1, 1, 0, 0, tzinfo=timezone.utc),
                 provider_name='openai',
                 provider_request_id='123',
@@ -155,18 +161,28 @@ async def test_request_simple_success(allow_model_requests: None):
     ]
 
 
-async def test_request_simple_usage(allow_model_requests: None):
-    c = completion_message(
-        ChatCompletionMessage(content='world', role='assistant'),
-        usage=CompletionUsage(completion_tokens=1, prompt_tokens=2, total_tokens=3),
-    )
-    mock_client = MockOpenAI.create_mock(c)
-    m = OpenAIChatModel('gpt-4o', provider=OpenAIProvider(openai_client=mock_client))
+async def test_request_simple_usage(allow_model_requests: None, openai_api_key: str):
+    m = OpenAIChatModel('gpt-4o', provider=OpenAIProvider(api_key=openai_api_key))
     agent = Agent(m)
 
-    result = await agent.run('Hello')
-    assert result.output == 'world'
-    assert result.usage() == snapshot(RunUsage(requests=1, input_tokens=2, output_tokens=1))
+    result = await agent.run('Hello! How are you doing?')
+    assert result.output == snapshot(
+        "Hello! I'm just a computer program, so I don't have feelings, but I'm here and ready to help you. How can I assist you today?"
+    )
+    assert result.usage() == snapshot(
+        RunUsage(
+            requests=1,
+            input_tokens=14,
+            details={
+                'accepted_prediction_tokens': 0,
+                'audio_tokens': 0,
+                'reasoning_tokens': 0,
+                'rejected_prediction_tokens': 0,
+            },
+            output_tokens=30,
+            cost=Decimal('0.000335'),
+        )
+    )
 
 
 async def test_request_structured_response(allow_model_requests: None):
@@ -200,7 +216,7 @@ async def test_request_structured_response(allow_model_requests: None):
                         tool_call_id='123',
                     )
                 ],
-                model_name='gpt-4o-123',
+                model_name='gpt-4o',
                 timestamp=datetime(2024, 1, 1, tzinfo=timezone.utc),
                 provider_name='openai',
                 provider_request_id='123',
@@ -295,7 +311,7 @@ async def get_location(loc_name: str) -> str:
                     cache_read_tokens=1,
                     output_tokens=1,
                 ),
-                model_name='gpt-4o-123',
+                model_name='gpt-4o',
                 timestamp=datetime(2024, 1, 1, tzinfo=timezone.utc),
                 provider_name='openai',
                 provider_request_id='123',
@@ -323,7 +339,7 @@ async def get_location(loc_name: str) -> str:
                     cache_read_tokens=2,
                     output_tokens=2,
                 ),
-                model_name='gpt-4o-123',
+                model_name='gpt-4o',
                 timestamp=datetime(2024, 1, 1, tzinfo=timezone.utc),
                 provider_name='openai',
                 provider_request_id='123',
@@ -340,14 +356,16 @@ async def get_location(loc_name: str) -> str:
             ),
             ModelResponse(
                 parts=[TextPart(content='final response')],
-                model_name='gpt-4o-123',
+                model_name='gpt-4o',
                 timestamp=datetime(2024, 1, 1, tzinfo=timezone.utc),
                 provider_name='openai',
                 provider_request_id='123',
             ),
         ]
     )
-    assert result.usage() == snapshot(RunUsage(requests=3, cache_read_tokens=3, input_tokens=5, output_tokens=3))
+    assert result.usage() == snapshot(
+        RunUsage(requests=3, cache_read_tokens=3, input_tokens=5, output_tokens=3, cost=Decimal('0.00004625'))
+    )
 
 
 FinishReason = Literal['stop', 'length', 'tool_calls', 'content_filter', 'function_call']
@@ -2273,6 +2291,8 @@ def test_model_profile_strict_not_supported():
     )
 
 
+# NOTE(Marcelo): You wouldn't do this because you'd use the GoogleModel. I'm unsure if this test brings any value.
+@pytest.mark.filterwarnings('ignore:The costs with provider "openai" and model "gemini-2.5-pro:UserWarning')
 async def test_compatible_api_with_tool_calls_without_id(allow_model_requests: None, gemini_api_key: str):
     provider = OpenAIProvider(
         openai_client=AsyncOpenAI(

tests/models/test_openai_responses.py

@@ -1115,7 +1115,7 @@ async def test_openai_responses_usage_without_tokens_details(allow_model_request
             ModelResponse(
                 parts=[TextPart(content='4')],
                 usage=RequestUsage(input_tokens=14, output_tokens=1, details={'reasoning_tokens': 0}),
-                model_name='gpt-4o-123',
+                model_name='gpt-4o',
                 timestamp=IsDatetime(),
                 provider_name='openai',
                 provider_request_id='123',