Skip to content

The agent engineering platform.

openai[patch]: add usage metadata details (#27080)

Bagaturc09da539788e
diff --git a/libs/partners/openai/langchain_openai/chat_models/base.py b/libs/partners/openai/langchain_openai/chat_models/base.py
index 21754937d1..27d4adf068 100644
--- a/libs/partners/openai/langchain_openai/chat_models/base.py
+++ b/libs/partners/openai/langchain_openai/chat_models/base.py
@@ -63,7 +63,11 @@ from langchain_core.messages import (
ToolMessage,
ToolMessageChunk,
)
-from langchain_core.messages.ai import UsageMetadata
+from langchain_core.messages.ai import (
+ InputTokenDetails,
+ OutputTokenDetails,
+ UsageMetadata,
+)
from langchain_core.messages.tool import tool_call_chunk
from langchain_core.output_parsers import JsonOutputParser, PydanticOutputParser
from langchain_core.output_parsers.openai_tools import (
@@ -286,16 +290,10 @@ def _convert_chunk_to_generation_chunk(
) -> Optional[ChatGenerationChunk]:
token_usage = chunk.get("usage")
choices = chunk.get("choices", [])
+
usage_metadata: Optional[UsageMetadata] = (
- UsageMetadata(
- input_tokens=token_usage.get("prompt_tokens", 0),
- output_tokens=token_usage.get("completion_tokens", 0),
- total_tokens=token_usage.get("total_tokens", 0),
- )
- if token_usage
- else None
+ _create_usage_metadata(token_usage) if token_usage else None
)
-
if len(choices) == 0:
# logprobs is implicitly None
generation_chunk = ChatGenerationChunk(
@@ -721,15 +719,11 @@ class BaseChatOpenAI(BaseChatModel):
if response_dict.get("error"):
raise ValueError(response_dict.get("error"))
- token_usage = response_dict.get("usage", {})
+ token_usage = response_dict.get("usage")
for res in response_dict["choices"]:
message = _convert_dict_to_message(res["message"])
if token_usage and isinstance(message, AIMessage):
- message.usage_metadata = {
- "input_tokens": token_usage.get("prompt_tokens", 0),
- "output_tokens": token_usage.get("completion_tokens", 0),
- "total_tokens": token_usage.get("total_tokens", 0),
- }
+ message.usage_metadata = _create_usage_metadata(token_usage)
generation_info = generation_info or {}
generation_info["finish_reason"] = (
res.get("finish_reason")
@@ -2160,3 +2154,34 @@ class OpenAIRefusalError(Exception):
.. versionadded:: 0.1.21
"""
+
+
+def _create_usage_metadata(oai_token_usage: dict) -> UsageMetadata:
+ input_tokens = oai_token_usage.get("prompt_tokens", 0)
+ output_tokens = oai_token_usage.get("completion_tokens", 0)
+ total_tokens = oai_token_usage.get("total_tokens", input_tokens + output_tokens)
+ input_token_details: dict = {
+ "audio": oai_token_usage.get("prompt_tokens_details", {}).get("audio_tokens"),
+ "cache_read": oai_token_usage.get("prompt_tokens_details", {}).get(
+ "cached_tokens"
+ ),
+ }
+ output_token_details: dict = {
+ "audio": oai_token_usage.get("completion_tokens_details", {}).get(
+ "audio_tokens"
+ ),
+ "reasoning": oai_token_usage.get("completion_tokens_details", {}).get(
+ "reasoning_tokens"
+ ),
+ }
+ return UsageMetadata(
+ input_tokens=input_tokens,
+ output_tokens=output_tokens,
+ total_tokens=total_tokens,
+ input_token_details=InputTokenDetails(
+ **{k: v for k, v in input_token_details.items() if v is not None}
+ ),
+ output_token_details=OutputTokenDetails(
+ **{k: v for k, v in output_token_details.items() if v is not None}
+ ),
+ )
diff --git a/libs/partners/openai/tests/integration_tests/chat_models/test_base_standard.py b/libs/partners/openai/tests/integration_tests/chat_models/test_base_standard.py
index 22b305c975..b91b590ad7 100644
--- a/libs/partners/openai/tests/integration_tests/chat_models/test_base_standard.py
+++ b/libs/partners/openai/tests/integration_tests/chat_models/test_base_standard.py
@@ -1,12 +1,16 @@
"""Standard LangChain interface tests"""
-from typing import Type
+from pathlib import Path
+from typing import List, Literal, Type, cast
from langchain_core.language_models import BaseChatModel
+from langchain_core.messages import AIMessage
from langchain_standard_tests.integration_tests import ChatModelIntegrationTests
from langchain_openai import ChatOpenAI
+REPO_ROOT_DIR = Path(__file__).parents[6]
+
class TestOpenAIStandard(ChatModelIntegrationTests):
@property
@@ -20,3 +24,48 @@ class TestOpenAIStandard(ChatModelIntegrationTests):
@property
def supports_image_inputs(self) -> bool:
return True
+
+ @property
+ def supported_usage_metadata_details(
+ self,
+ ) -> List[
+ Literal[
+ "audio_input",
+ "audio_output",
+ "reasoning_output",
+ "cache_read_input",
+ "cache_creation_input",
+ ]
+ ]:
+ return ["reasoning_output", "cache_read_input"]
+
+ def invoke_with_cache_read_input(self, *, stream: bool = False) -> AIMessage:
+ with open(REPO_ROOT_DIR / "README.md", "r") as f:
+ readme = f.read()
+
+ input_ = f"""What's langchain? Here's the langchain README:
+
+ {readme}
+ """
+ llm = ChatOpenAI(model="gpt-4o-mini", stream_usage=True)
+ _invoke(llm, input_, stream)
+ # invoke twice so first invocation is cached
+ return _invoke(llm, input_, stream)
+
+ def invoke_with_reasoning_output(self, *, stream: bool = False) -> AIMessage:
+ llm = ChatOpenAI(model="o1-mini", stream_usage=True, temperature=1)
+ input_ = (
+ "explain the relationship between the 2008/9 economic crisis and the "
+ "startup ecosystem in the early 2010s"
+ )
+ return _invoke(llm, input_, stream)
+
+
+def _invoke(llm: ChatOpenAI, input_: str, stream: bool) -> AIMessage:
+ if stream:
+ full = None
+ for chunk in llm.stream(input_):
+ full = full + chunk if full else chunk # type: ignore[operator]
+ return cast(AIMessage, full)
+ else:
+ return cast(AIMessage, llm.invoke(input_))