Raise IncompleteToolCall when token limit is reached during generation of tool call

pydantic_ai_slim/pydantic_ai/__init__.py

@@ -24,6 +24,7 @@
     FallbackExceptionGroup,
     ModelHTTPError,
     ModelRetry,
+    ToolExceedsTokenLimitError,
     UnexpectedModelBehavior,
     UsageLimitExceeded,
     UserError,
@@ -124,6 +125,7 @@
     'ModelRetry',
     'ModelHTTPError',
     'FallbackExceptionGroup',
+    'ToolExceedsTokenLimitError',
     'UnexpectedModelBehavior',
     'UsageLimitExceeded',
     'UserError',

pydantic_ai_slim/pydantic_ai/_agent_graph.py

@@ -92,9 +92,24 @@ class GraphAgentState:
     retries: int = 0
     run_step: int = 0
 
-    def increment_retries(self, max_result_retries: int, error: BaseException | None = None) -> None:
+    def increment_retries(
+        self, max_result_retries: int, error: BaseException | None = None, max_tokens: int | None = None
+    ) -> None:
         self.retries += 1
         if self.retries > max_result_retries:
+            if (
+                self.message_history
+                and isinstance(model_response := self.message_history[-1], _messages.ModelResponse)
+                and model_response.finish_reason == 'length'
+                and model_response.parts
+                and isinstance(tool_call := model_response.parts[-1], _messages.ToolCallPart)
+            ):
+                try:
+                    tool_call.args_as_dict()
+                except Exception:
+                    raise exceptions.ToolExceedsTokenLimitError(
+                        f'Model token limit ({max_tokens if max_tokens is not None else "provider default"}) exceeded while emitting a tool call, resulting in incomplete arguments. Increase max tokens or simplify tool call arguments to fit within limit.'
+                    )
             message = f'Exceeded maximum retries ({max_result_retries}) for output validation'
             if error:
                 if isinstance(error, exceptions.UnexpectedModelBehavior) and error.__cause__ is not None:
@@ -568,7 +583,8 @@ async def _run_stream() -> AsyncIterator[_messages.HandleResponseEvent]:  # noqa
                     # resubmit the most recent request that resulted in an empty response,
                     # as the empty response and request will not create any items in the API payload,
                     # in the hope the model will return a non-empty response this time.
-                    ctx.state.increment_retries(ctx.deps.max_result_retries)
+                    max_tokens = (ctx.deps.model_settings or {}).get('max_tokens') if ctx.deps.model_settings else None
+                    ctx.state.increment_retries(ctx.deps.max_result_retries, max_tokens=max_tokens)
                     self._next_node = ModelRequestNode[DepsT, NodeRunEndT](_messages.ModelRequest(parts=[]))
                     return
 
@@ -630,7 +646,8 @@ async def _run_stream() -> AsyncIterator[_messages.HandleResponseEvent]:  # noqa
                     )
                     raise ToolRetryError(m)
                 except ToolRetryError as e:
-                    ctx.state.increment_retries(ctx.deps.max_result_retries, e)
+                    max_tokens = (ctx.deps.model_settings or {}).get('max_tokens') if ctx.deps.model_settings else None
+                    ctx.state.increment_retries(ctx.deps.max_result_retries, error=e, max_tokens=max_tokens)
                     self._next_node = ModelRequestNode[DepsT, NodeRunEndT](_messages.ModelRequest(parts=[e.tool_retry]))
 
             self._events_iterator = _run_stream()
@@ -776,10 +793,12 @@ async def process_tool_calls(  # noqa: C901
             try:
                 result_data = await tool_manager.handle_call(call)
             except exceptions.UnexpectedModelBehavior as e:
-                ctx.state.increment_retries(ctx.deps.max_result_retries, e)
+                max_tokens = (ctx.deps.model_settings or {}).get('max_tokens') if ctx.deps.model_settings else None
+                ctx.state.increment_retries(ctx.deps.max_result_retries, error=e, max_tokens=max_tokens)
                 raise e  # pragma: lax no cover
             except ToolRetryError as e:
-                ctx.state.increment_retries(ctx.deps.max_result_retries, e)
+                max_tokens = (ctx.deps.model_settings or {}).get('max_tokens') if ctx.deps.model_settings else None
+                ctx.state.increment_retries(ctx.deps.max_result_retries, error=e, max_tokens=max_tokens)
                 yield _messages.FunctionToolCallEvent(call)
                 output_parts.append(e.tool_retry)
                 yield _messages.FunctionToolResultEvent(e.tool_retry)
@@ -808,7 +827,8 @@ async def process_tool_calls(  # noqa: C901
 
     # Then, we handle unknown tool calls
     if tool_calls_by_kind['unknown']:
-        ctx.state.increment_retries(ctx.deps.max_result_retries)
+        max_tokens = (ctx.deps.model_settings or {}).get('max_tokens') if ctx.deps.model_settings else None
+        ctx.state.increment_retries(ctx.deps.max_result_retries, max_tokens=max_tokens)
         calls_to_run.extend(tool_calls_by_kind['unknown'])
 
     calls_to_run_results: dict[str, DeferredToolResult] = {}

pydantic_ai_slim/pydantic_ai/exceptions.py

@@ -23,6 +23,7 @@
     'UnexpectedModelBehavior',
     'UsageLimitExceeded',
     'ModelHTTPError',
+    'ToolExceedsTokenLimitError',
     'FallbackExceptionGroup',
 )
 
@@ -168,3 +169,7 @@ class ToolRetryError(Exception):
     def __init__(self, tool_retry: RetryPromptPart):
         self.tool_retry = tool_retry
         super().__init__()
+
+
+class ToolExceedsTokenLimitError(AgentRunError):
+    """Error raised when a model stops due to token limit while emitting a tool call."""

tests/test_agent.py

@@ -40,6 +40,7 @@
     SystemPromptPart,
     TextPart,
     ToolCallPart,
+    ToolExceedsTokenLimitError,
     ToolReturn,
     ToolReturnPart,
     UnexpectedModelBehavior,
@@ -63,6 +64,7 @@
 from pydantic_ai.output import StructuredDict, ToolOutput
 from pydantic_ai.result import RunUsage
 from pydantic_ai.run import AgentRunResultEvent
+from pydantic_ai.settings import ModelSettings
 from pydantic_ai.tools import DeferredToolRequests, DeferredToolResults, ToolDefinition, ToolDenied
 from pydantic_ai.usage import RequestUsage
 
@@ -2448,6 +2450,45 @@ def empty(m: list[ModelMessage], _info: AgentInfo) -> ModelResponse:
     )
 
 
+def test_tool_exceeds_token_limit_error():
+    def return_incomplete_tool(_: list[ModelMessage], info: AgentInfo) -> ModelResponse:
+        resp = ModelResponse(parts=[ToolCallPart('dummy_tool', args='{"foo": "bar",')])
+        resp.finish_reason = 'length'
+        return resp
+
+    agent = Agent(FunctionModel(return_incomplete_tool), output_type=str)
+
+    with pytest.raises(
+        ToolExceedsTokenLimitError,
+        match=r'Model token limit \(10\) exceeded while emitting a tool call, resulting in incomplete arguments. Increase max tokens or simplify tool call arguments to fit within limit.',
+    ):
+        agent.run_sync('Hello', model_settings=ModelSettings(max_tokens=10))
+
+    with pytest.raises(
+        ToolExceedsTokenLimitError,
+        match=r'Model token limit \(provider default\) exceeded while emitting a tool call, resulting in incomplete arguments. Increase max tokens or simplify tool call arguments to fit within limit.',
+    ):
+        agent.run_sync('Hello')
+
+
+def test_tool_exceeds_token_limit_but_complete_args():
+    def return_complete_tool_but_hit_limit(messages: list[ModelMessage], info: AgentInfo) -> ModelResponse:
+        if len(messages) == 1:
+            resp = ModelResponse(parts=[ToolCallPart('dummy_tool', args='{"foo": "bar"}')])
+            resp.finish_reason = 'length'
+            return resp
+        return ModelResponse(parts=[TextPart('done')])
+
+    agent = Agent(FunctionModel(return_complete_tool_but_hit_limit), output_type=str)
+
+    @agent.tool_plain
+    def dummy_tool(foo: str) -> str:
+        return 'tool-ok'
+
+    result = agent.run_sync('Hello')
+    assert result.output == 'done'
+
+
 def test_model_requests_blocked(env: TestEnv):
     try:
         env.set('GEMINI_API_KEY', 'foobar')