langchain-ai/langgraph
1
0
Fork 0
mirror of https://github.com/langchain-ai/langgraph.git synced 2026-07-15 12:35:43 -04:00
Code Issues 314 Packages Projects Releases 17 Wiki Activity

Gemini 2.5 Fails with LangGraph Agent #638

New Issue
Closed
opened 2026-02-20 17:41:04 -05:00 by yindo · 14 comments
yindo commented 2026-02-20 17:41:04 -05:00
Owner
Copy Link
Copy Source

Originally created by @jiveshkalra on GitHub (May 21, 2025).

Checked other resources

  • This is a bug, not a usage question. For questions, please use GitHub Discussions.
  • I added a clear and detailed title that summarizes the issue.
  • I read what a minimal reproducible example is (https://stackoverflow.com/help/minimal-reproducible-example).
  • I included a self-contained, minimal example that demonstrates the issue INCLUDING all the relevant imports. The code run AS IS to reproduce the issue.

Example Code

from dotenv import load_dotenv
import os
from langchain.chat_models import init_chat_model
from typing import Annotated
from langchain_tavily import TavilySearch
from langchain_core.messages import BaseMessage
from typing_extensions import TypedDict
from langgraph.graph import StateGraph, START, END
from langgraph.graph.message import add_messages
from langgraph.prebuilt import ToolNode, tools_condition

load_dotenv()
TAVILY_API_KEY = os.getenv("TAVILY_API_KEY")  
GEMINI_API_KEY = os.getenv("GEMINI_API_KEY") 

llm = init_chat_model("google_genai:gemini-2.5-flash-preview-05-20")

class State(TypedDict):
    messages: Annotated[list, add_messages]

graph_builder = StateGraph(State)

tool = TavilySearch(max_results=2)
tools = [tool]
llm_with_tools = llm.bind_tools(tools)

def chatbot(state: State):
    return {"messages": [llm_with_tools.invoke(state["messages"])]}


graph_builder.add_node("chatbot", chatbot)

tool_node = ToolNode(tools=[tool])

graph_builder.add_node("tools", tool_node)

graph_builder.add_edge(START, "chatbot")
graph_builder.add_conditional_edges(
    "chatbot",
    tools_condition,
)
# Any time a tool is called, we return to the chatbot to decide the next step
graph_builder.add_edge("tools", "chatbot")
graph = graph_builder.compile()

def stream_graph_updates(user_input: str):
    for event in graph.stream({"messages": [{"role": "user", "content": user_input}]}):
        for value in event.values():
            print("Assistant:", value["messages"][-1].content)


user_input = "Tell me the weather in New York City today. "

stream_graph_updates(user_input)

Error Message and Stack Trace (if applicable)

Gemini produced an empty response. Continuing with empty message
Feedback: block_reason: OTHER

More detailed error : 
ChatGoogleGenerativeAIError: Invalid argument provided to Gemini: 400 * GenerateContentRequest.contents[2].parts: contents.parts must not be empty.

Description

I have a minimal working example of a LangGraph-based agent that uses Gemini 2.0 with tools via langchain. The setup works perfectly fine with Gemini 2.0 models. However, when switching to the Gemini 2.5 reasoning models (gemini-2.5-flash-preview-05-20), the agent fails—especially during tool invocation.

Expected Behavior

When using Gemini 2.5 with tool-binding in a LangGraph agent:

  • The agent should respond intelligently.
  • It should invoke tools like TavilySearch when needed.
  • It should stream valid message content in response.

Actual Behavior

Instead of producing meaningful output, the agent crashes with the following:

Gemini produced an empty response. Continuing with empty message
block_reason: OTHER

ChatGoogleGenerativeAIError: Invalid argument provided to Gemini: 400 * GenerateContentRequest.contents[2].parts: contents.parts must not be empty.

System Info

System Information

OS: Windows
OS Version: 10.0.19045
Python Version: 3.12.7 | packaged by Anaconda, Inc. | (main, Oct 4 2024, 13:17:27) [MSC v.1929 64 bit (AMD64)]

Package Information

langchain_core: 0.3.59
langchain: 0.3.25
langsmith: 0.3.42
langchain_google_genai: 2.1.4
langchain_tavily: 0.1.6
langchain_text_splitters: 0.3.8
langgraph_sdk: 0.1.69

Optional packages not installed

langserve

Other Dependencies

aiohttp: 3.11.18
async-timeout<5.0.0,>=4.0.0;: Installed. No version info available.
filetype: 1.2.0
google-ai-generativelanguage: 0.6.15
httpx: 0.28.1
jsonpatch<2.0,>=1.33: Installed. No version info available.
langchain-anthropic;: Installed. No version info available.
langchain-aws;: Installed. No version info available.
langchain-azure-ai;: Installed. No version info available.
langchain-cohere;: Installed. No version info available.
langchain-community;: Installed. No version info available.
langchain-core<1.0.0,>=0.3.51: Installed. No version info available.
langchain-core<1.0.0,>=0.3.58: Installed. No version info available.
langchain-deepseek;: Installed. No version info available.
langchain-fireworks;: Installed. No version info available.
langchain-google-genai;: Installed. No version info available.
langchain-google-vertexai;: Installed. No version info available.
langchain-groq;: Installed. No version info available.
langchain-huggingface;: Installed. No version info available.
langchain-mistralai;: Installed. No version info available.
langchain-ollama;: Installed. No version info available.
langchain-openai;: Installed. No version info available.
langchain-perplexity;: Installed. No version info available.
langchain-text-splitters<1.0.0,>=0.3.8: Installed. No version info available.
langchain-together;: Installed. No version info available.
langchain-xai;: Installed. No version info available.
langsmith-pyo3: Installed. No version info available.
langsmith<0.4,>=0.1.125: Installed. No version info available.
langsmith<0.4,>=0.1.17: Installed. No version info available.
mypy: 1.15.0
openai-agents: Installed. No version info available.
opentelemetry-api: 1.33.0
opentelemetry-exporter-otlp-proto-http: 1.33.0
opentelemetry-sdk: 1.33.0
orjson: 3.10.18
packaging: 24.1
packaging<25,>=23.2: Installed. No version info available.
pydantic: 2.11.4
pydantic<3.0.0,>=2.5.2;: Installed. No version info available.
pydantic<3.0.0,>=2.7.4: Installed. No version info available.
pydantic<3.0.0,>=2.7.4;: Installed. No version info available.
pytest: 7.4.4
PyYAML>=5.3: Installed. No version info available.
requests: 2.32.3
requests-toolbelt: 1.0.0
requests<3,>=2: Installed. No version info available.
rich: 13.9.4
SQLAlchemy<3,>=1.4: Installed. No version info available.
tenacity!=8.4.0,<10.0.0,>=8.1.0: Installed. No version info available.
typing-extensions>=4.7: Installed. No version info available.
zstandard: 0.23.0

Originally created by @jiveshkalra on GitHub (May 21, 2025). ### Checked other resources - [x] This is a bug, not a usage question. For questions, please use GitHub Discussions. - [x] I added a clear and detailed title that summarizes the issue. - [x] I read what a minimal reproducible example is (https://stackoverflow.com/help/minimal-reproducible-example). - [x] I included a self-contained, minimal example that demonstrates the issue INCLUDING all the relevant imports. The code run AS IS to reproduce the issue. ### Example Code ```python from dotenv import load_dotenv import os from langchain.chat_models import init_chat_model from typing import Annotated from langchain_tavily import TavilySearch from langchain_core.messages import BaseMessage from typing_extensions import TypedDict from langgraph.graph import StateGraph, START, END from langgraph.graph.message import add_messages from langgraph.prebuilt import ToolNode, tools_condition load_dotenv() TAVILY_API_KEY = os.getenv("TAVILY_API_KEY") GEMINI_API_KEY = os.getenv("GEMINI_API_KEY") llm = init_chat_model("google_genai:gemini-2.5-flash-preview-05-20") class State(TypedDict): messages: Annotated[list, add_messages] graph_builder = StateGraph(State) tool = TavilySearch(max_results=2) tools = [tool] llm_with_tools = llm.bind_tools(tools) def chatbot(state: State): return {"messages": [llm_with_tools.invoke(state["messages"])]} graph_builder.add_node("chatbot", chatbot) tool_node = ToolNode(tools=[tool]) graph_builder.add_node("tools", tool_node) graph_builder.add_edge(START, "chatbot") graph_builder.add_conditional_edges( "chatbot", tools_condition, ) # Any time a tool is called, we return to the chatbot to decide the next step graph_builder.add_edge("tools", "chatbot") graph = graph_builder.compile() def stream_graph_updates(user_input: str): for event in graph.stream({"messages": [{"role": "user", "content": user_input}]}): for value in event.values(): print("Assistant:", value["messages"][-1].content) user_input = "Tell me the weather in New York City today. " stream_graph_updates(user_input) ``` ### Error Message and Stack Trace (if applicable) ```shell Gemini produced an empty response. Continuing with empty message Feedback: block_reason: OTHER More detailed error : ChatGoogleGenerativeAIError: Invalid argument provided to Gemini: 400 * GenerateContentRequest.contents[2].parts: contents.parts must not be empty. ``` ### Description I have a minimal working example of a LangGraph-based agent that uses `Gemini 2.0` with tools via langchain. The setup works perfectly fine with Gemini 2.0 models. However, when switching to the `Gemini 2.5` **reasoning** models (gemini-2.5-flash-preview-05-20), the agent fails—especially during tool invocation. ## Expected Behavior When using Gemini 2.5 with tool-binding in a LangGraph agent: - The agent should respond intelligently. - It should invoke tools like TavilySearch when needed. - It should stream valid message content in response. ## Actual Behavior Instead of producing meaningful output, the agent crashes with the following: ``` Gemini produced an empty response. Continuing with empty message block_reason: OTHER ChatGoogleGenerativeAIError: Invalid argument provided to Gemini: 400 * GenerateContentRequest.contents[2].parts: contents.parts must not be empty. ``` ### System Info System Information ------------------ > OS: Windows > OS Version: 10.0.19045 > Python Version: 3.12.7 | packaged by Anaconda, Inc. | (main, Oct 4 2024, 13:17:27) [MSC v.1929 64 bit (AMD64)] Package Information ------------------- > langchain_core: 0.3.59 > langchain: 0.3.25 > langsmith: 0.3.42 > langchain_google_genai: 2.1.4 > langchain_tavily: 0.1.6 > langchain_text_splitters: 0.3.8 > langgraph_sdk: 0.1.69 Optional packages not installed ------------------------------- > langserve Other Dependencies ------------------ > aiohttp: 3.11.18 > async-timeout<5.0.0,>=4.0.0;: Installed. No version info available. > filetype: 1.2.0 > google-ai-generativelanguage: 0.6.15 > httpx: 0.28.1 > jsonpatch<2.0,>=1.33: Installed. No version info available. > langchain-anthropic;: Installed. No version info available. > langchain-aws;: Installed. No version info available. > langchain-azure-ai;: Installed. No version info available. > langchain-cohere;: Installed. No version info available. > langchain-community;: Installed. No version info available. > langchain-core<1.0.0,>=0.3.51: Installed. No version info available. > langchain-core<1.0.0,>=0.3.58: Installed. No version info available. > langchain-deepseek;: Installed. No version info available. > langchain-fireworks;: Installed. No version info available. > langchain-google-genai;: Installed. No version info available. > langchain-google-vertexai;: Installed. No version info available. > langchain-groq;: Installed. No version info available. > langchain-huggingface;: Installed. No version info available. > langchain-mistralai;: Installed. No version info available. > langchain-ollama;: Installed. No version info available. > langchain-openai;: Installed. No version info available. > langchain-perplexity;: Installed. No version info available. > langchain-text-splitters<1.0.0,>=0.3.8: Installed. No version info available. > langchain-together;: Installed. No version info available. > langchain-xai;: Installed. No version info available. > langsmith-pyo3: Installed. No version info available. > langsmith<0.4,>=0.1.125: Installed. No version info available. > langsmith<0.4,>=0.1.17: Installed. No version info available. > mypy: 1.15.0 > openai-agents: Installed. No version info available. > opentelemetry-api: 1.33.0 > opentelemetry-exporter-otlp-proto-http: 1.33.0 > opentelemetry-sdk: 1.33.0 > orjson: 3.10.18 > packaging: 24.1 > packaging<25,>=23.2: Installed. No version info available. > pydantic: 2.11.4 > pydantic<3.0.0,>=2.5.2;: Installed. No version info available. > pydantic<3.0.0,>=2.7.4: Installed. No version info available. > pydantic<3.0.0,>=2.7.4;: Installed. No version info available. > pytest: 7.4.4 > PyYAML>=5.3: Installed. No version info available. > requests: 2.32.3 > requests-toolbelt: 1.0.0 > requests<3,>=2: Installed. No version info available. > rich: 13.9.4 > SQLAlchemy<3,>=1.4: Installed. No version info available. > tenacity!=8.4.0,<10.0.0,>=8.1.0: Installed. No version info available. > typing-extensions>=4.7: Installed. No version info available. > zstandard: 0.23.0
yindo added the bug label 2026-02-20 17:41:04 -05:00
yindo closed this issue 2026-02-20 17:41:04 -05:00
yindo commented 2026-02-20 17:41:04 -05:00
Author
Owner
Copy Link
Copy Source

@jiveshkalra commented on GitHub (May 21, 2025):

PS: Even the prebuilt React agents dont work with gemini
I'm getting same error when using this

from langgraph.prebuilt import create_react_agent
from dotenv import load_dotenv
import os
 
load_dotenv()

TAVILY_API_KEY = os.getenv("TAVILY_API_KEY")  
GEMINI_API_KEY = os.getenv("GEMINI_API_KEY")
os.environ["GOOGLE_API_KEY"] = GEMINI_API_KEY 

def get_weather(city: str) -> str:
    """Get weather for a given city."""
    return f"It's always sunny in {city}!"

agent = create_react_agent(
    model="google_genai:gemini-2.5-flash-preview-05-20",
    tools=[get_weather],
    prompt="You are a helpful assistant"
)
 
agent.invoke(
    {"messages": [{"role": "user", "content": "what is the weather in sf"}]}
)

Error :

Gemini produced an empty response. Continuing with empty message
Feedback: block_reason: OTHER

@jiveshkalra commented on GitHub (May 21, 2025): PS: Even the prebuilt React agents dont work with gemini I'm getting same error when using this ```py from langgraph.prebuilt import create_react_agent from dotenv import load_dotenv import os load_dotenv() TAVILY_API_KEY = os.getenv("TAVILY_API_KEY") GEMINI_API_KEY = os.getenv("GEMINI_API_KEY") os.environ["GOOGLE_API_KEY"] = GEMINI_API_KEY def get_weather(city: str) -> str: """Get weather for a given city.""" return f"It's always sunny in {city}!" agent = create_react_agent( model="google_genai:gemini-2.5-flash-preview-05-20", tools=[get_weather], prompt="You are a helpful assistant" ) agent.invoke( {"messages": [{"role": "user", "content": "what is the weather in sf"}]} ) ``` ## Error : Gemini produced an empty response. Continuing with empty message Feedback: block_reason: OTHER
yindo commented 2026-02-20 17:41:04 -05:00
Author
Owner
Copy Link
Copy Source

@sergey-koba-mobidev commented on GitHub (May 23, 2025):

https://github.com/langchain-ai/langchain-google/issues/936

@sergey-koba-mobidev commented on GitHub (May 23, 2025): https://github.com/langchain-ai/langchain-google/issues/936
yindo commented 2026-02-20 17:41:04 -05:00
Author
Owner
Copy Link
Copy Source

@sydney-runkle commented on GitHub (May 28, 2025):

Seems based on https://github.com/langchain-ai/langchain-google/issues/936#issuecomment-2911706467 that this is resolved on the google end - I'm no longer able to reproduce.

Going to close as resolved, feel free to ping me if you are still experiencing this problem.

@sydney-runkle commented on GitHub (May 28, 2025): Seems based on https://github.com/langchain-ai/langchain-google/issues/936#issuecomment-2911706467 that this is resolved on the google end - I'm no longer able to reproduce. Going to close as resolved, feel free to ping me if you are still experiencing this problem.
yindo commented 2026-02-20 17:41:04 -05:00
Author
Owner
Copy Link
Copy Source

@xajik commented on GitHub (Jun 1, 2025):

I can reproduce it 100% when I use tools with the Gemini model:

Works:
runnable = prompt | self._llm
Doesn't work
runnable = prompt | self._llm.bind_tools(self._tools)

Tools are coming from my MCP.

langchain_google_genai.chat_models.ChatGoogleGenerativeAIError: Invalid argument provided to Gemini: 400 * GenerateContentRequest.contents[2].parts: contents.parts must not be empty.
@xajik commented on GitHub (Jun 1, 2025): I can reproduce it 100% when I use tools with the Gemini model: Works: `runnable = prompt | self._llm` Doesn't work `runnable = prompt | self._llm.bind_tools(self._tools)` Tools are coming from my MCP. ``` langchain_google_genai.chat_models.ChatGoogleGenerativeAIError: Invalid argument provided to Gemini: 400 * GenerateContentRequest.contents[2].parts: contents.parts must not be empty. ```
yindo commented 2026-02-20 17:41:04 -05:00
Author
Owner
Copy Link
Copy Source

@jiveshkalra commented on GitHub (Jun 5, 2025):

same here , it is still not fixed

@jiveshkalra commented on GitHub (Jun 5, 2025): same here , it is still not fixed
yindo commented 2026-02-20 17:41:04 -05:00
Author
Owner
Copy Link
Copy Source

@codykingham commented on GitHub (Jun 10, 2025):

This issue is not solved.

@codykingham commented on GitHub (Jun 10, 2025): This issue is not solved.
yindo commented 2026-02-20 17:41:04 -05:00
Author
Owner
Copy Link
Copy Source

@jiveshkalra commented on GitHub (Jun 12, 2025):

Anyone found any hack/trick to work around it?
I think it might be issue in the google's own genAI kit , had same issue with that library

@jiveshkalra commented on GitHub (Jun 12, 2025): Anyone found any hack/trick to work around it? I think it might be issue in the google's own genAI kit , had same issue with that library
yindo commented 2026-02-20 17:41:05 -05:00
Author
Owner
Copy Link
Copy Source

@Talha-Tahir-Khurshid commented on GitHub (Jun 20, 2025):

Gemini has a problem where it does not work if text part of a tool call is empty , so it throws an error , I hacked around it by adding No content as str to my tool calls before passing them to the Gemini Model. This is google problem.

you can use a similar function to this

def sanitize_message_content(messages: List[BaseMessage]) -> List[BaseMessage]:
"""
Clean messages by removing empty text entries from content arrays to prevent
Gemini API errors related to empty text parameters.

Args:
    messages: List of BaseMessage objects to sanitize
    
Returns:
    List[BaseMessage]: Sanitized message list with empty text entries removed
"""
sanitized_messages = []

for msg in messages:
    # Create a copy of the message to avoid modifying the original
    if isinstance(msg, AIMessage) or isinstance(msg, AIMessageChunk):
        # Handle content that is a list of content items (dict with type/text)
        if isinstance(msg.content, list):
            # Filter out content items with empty text
            non_empty_content = [
                item for item in msg.content 
                if not (item.get('type') == 'text' and (item.get('text', '') == '' or item.get('text') is None))
            ]
            
            # If we have non-empty content, create a new message with it
            if non_empty_content:
                new_msg = AIMessage(
                    content=non_empty_content,
                    additional_kwargs=msg.additional_kwargs,
                    response_metadata=msg.response_metadata,
                    id=msg.id,
                    tool_calls=getattr(msg, 'tool_calls', None),
                    usage_metadata=getattr(msg, 'usage_metadata', None)
                )
                sanitized_messages.append(new_msg)
            else:
                # If all content was empty and removed, create a simple non-empty message
                # to maintain the conversation flow
                new_msg = AIMessage(
                    content=[{'type': 'text', 'text': '[No content]'}],
                    additional_kwargs=msg.additional_kwargs,
                    response_metadata=msg.response_metadata,
                    id=msg.id,
                    tool_calls=getattr(msg, 'tool_calls', None),
                    usage_metadata=getattr(msg, 'usage_metadata', None)
                )
                sanitized_messages.append(new_msg)
        else:
            # Handle string content - ensure it's not empty
            if not msg.content:
                content = "[No content]"
            # Handle case when content is a string representation of a list with empty text
            elif isinstance(msg.content, str) and msg.content.startswith("[{'type': 'text', 'text': ") and ("''" in msg.content or '""' in msg.content):
                content = "[No content]"
            elif isinstance(msg.content, list) and len(msg.content) == 1:
                item = msg.content[0]
                if isinstance(item, dict) and item.get('type') == 'text' and not item.get('text', '').strip():
                    content = "[No content]"
                else:
                    content = msg.content
            else:
                content = msg.content
            
            new_msg = AIMessage(
                content=[{'type': 'text', 'text': content}] if isinstance(content, str) else content,
                additional_kwargs=msg.additional_kwargs,
                response_metadata=msg.response_metadata,
                id=msg.id,
                model_fields_set=msg.model_fields_set,
                tool_calls=getattr(msg, 'tool_calls', None),
                usage_metadata=getattr(msg, 'usage_metadata', None)
            )
            sanitized_messages.append(new_msg)
    else:
        # For non-AI messages, keep as is
        sanitized_messages.append(msg)

return sanitized_messages
@Talha-Tahir-Khurshid commented on GitHub (Jun 20, 2025): Gemini has a problem where it does not work if text part of a tool call is empty , so it throws an error , I hacked around it by adding No content as str to my tool calls before passing them to the Gemini Model. This is google problem. you can use a similar function to this def sanitize_message_content(messages: List[BaseMessage]) -> List[BaseMessage]: """ Clean messages by removing empty text entries from content arrays to prevent Gemini API errors related to empty text parameters. Args: messages: List of BaseMessage objects to sanitize Returns: List[BaseMessage]: Sanitized message list with empty text entries removed """ sanitized_messages = [] for msg in messages: # Create a copy of the message to avoid modifying the original if isinstance(msg, AIMessage) or isinstance(msg, AIMessageChunk): # Handle content that is a list of content items (dict with type/text) if isinstance(msg.content, list): # Filter out content items with empty text non_empty_content = [ item for item in msg.content if not (item.get('type') == 'text' and (item.get('text', '') == '' or item.get('text') is None)) ] # If we have non-empty content, create a new message with it if non_empty_content: new_msg = AIMessage( content=non_empty_content, additional_kwargs=msg.additional_kwargs, response_metadata=msg.response_metadata, id=msg.id, tool_calls=getattr(msg, 'tool_calls', None), usage_metadata=getattr(msg, 'usage_metadata', None) ) sanitized_messages.append(new_msg) else: # If all content was empty and removed, create a simple non-empty message # to maintain the conversation flow new_msg = AIMessage( content=[{'type': 'text', 'text': '[No content]'}], additional_kwargs=msg.additional_kwargs, response_metadata=msg.response_metadata, id=msg.id, tool_calls=getattr(msg, 'tool_calls', None), usage_metadata=getattr(msg, 'usage_metadata', None) ) sanitized_messages.append(new_msg) else: # Handle string content - ensure it's not empty if not msg.content: content = "[No content]" # Handle case when content is a string representation of a list with empty text elif isinstance(msg.content, str) and msg.content.startswith("[{'type': 'text', 'text': ") and ("''" in msg.content or '""' in msg.content): content = "[No content]" elif isinstance(msg.content, list) and len(msg.content) == 1: item = msg.content[0] if isinstance(item, dict) and item.get('type') == 'text' and not item.get('text', '').strip(): content = "[No content]" else: content = msg.content else: content = msg.content new_msg = AIMessage( content=[{'type': 'text', 'text': content}] if isinstance(content, str) else content, additional_kwargs=msg.additional_kwargs, response_metadata=msg.response_metadata, id=msg.id, model_fields_set=msg.model_fields_set, tool_calls=getattr(msg, 'tool_calls', None), usage_metadata=getattr(msg, 'usage_metadata', None) ) sanitized_messages.append(new_msg) else: # For non-AI messages, keep as is sanitized_messages.append(msg) return sanitized_messages
yindo commented 2026-02-20 17:41:05 -05:00
Author
Owner
Copy Link
Copy Source

@lsm1041321457 commented on GitHub (Jun 25, 2025):

This problem have not been solved. I still encountered this problem and it's a fatal problem.

@lsm1041321457 commented on GitHub (Jun 25, 2025): This problem have not been solved. I still encountered this problem and it's a fatal problem.
yindo commented 2026-02-20 17:41:05 -05:00
Author
Owner
Copy Link
Copy Source

@tykimseoul commented on GitHub (Jun 26, 2025):

Not sure if this is the exact cause in your case, but I was able to resolve it by changing the first role from 'system' to 'user'.

generate_prompt = ChatPromptTemplate.from_messages(
    [
        ('user', generate_template), # was 'system'
        ('placeholder', '{messages}'),
    ]
)

Might be worth trying 🙂

@tykimseoul commented on GitHub (Jun 26, 2025): Not sure if this is the exact cause in your case, but I was able to resolve it by changing the first role from 'system' to 'user'. ```python generate_prompt = ChatPromptTemplate.from_messages( [ ('user', generate_template), # was 'system' ('placeholder', '{messages}'), ] ) ``` Might be worth trying 🙂
yindo commented 2026-02-20 17:41:05 -05:00
Author
Owner
Copy Link
Copy Source

@achugh-wald commented on GitHub (Jul 2, 2025):

It is surprising that such a critical issue has not been resolved for so long. Feels like a problem with VertexAI. Should this filed in some Google repo?

@achugh-wald commented on GitHub (Jul 2, 2025): It is surprising that such a critical issue has not been resolved for so long. Feels like a problem with VertexAI. Should this filed in some Google repo?
yindo commented 2026-02-20 17:41:05 -05:00
Author
Owner
Copy Link
Copy Source

@mohamed-em2m commented on GitHub (Jul 8, 2025):

try to not use bind to add stop using bind and add in initializee of main model dirctely
ChatOpenAI(
model="gpt-4.1-mini-2025-04-14",
api_key=os.environ["OPENAI_API_KEY"],stop="\nObservation:"
)

@mohamed-em2m commented on GitHub (Jul 8, 2025): try to not use bind to add stop using bind and add in initializee of main model dirctely ChatOpenAI( model="gpt-4.1-mini-2025-04-14", api_key=os.environ["OPENAI_API_KEY"],stop="\nObservation:" )
yindo commented 2026-02-20 17:41:05 -05:00
Author
Owner
Copy Link
Copy Source

@casparb commented on GitHub (Aug 13, 2025):

@lsm1041321457 @codykingham @jiveshkalra If you are still encountering this issue, could you please provide a minimal example? I am unable to reproduce this and it seems like Google have fixed this on their end. Looking to mark this as resolved.

@casparb commented on GitHub (Aug 13, 2025): @lsm1041321457 @codykingham @jiveshkalra If you are still encountering this issue, could you please provide a minimal example? I am unable to reproduce this and it seems like Google have fixed this on their end. Looking to mark this as resolved.
yindo commented 2026-02-20 17:41:05 -05:00
Author
Owner
Copy Link
Copy Source

@casparb commented on GitHub (Aug 14, 2025):

Closing as resolved. Please PM me if you are still experiencing this issue.

@casparb commented on GitHub (Aug 14, 2025): Closing as resolved. Please PM me if you are still experiencing this issue.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
agents
awaiting author revision
bug
change
chat models
checkpointer
ci
code quality
code quality
dependencies
documentation
drawing
duplicate
enhancement
good first issue
help wanted
invalid
maintainer
meta
missing information
not maintained by langchain
open-swe-auto
open-swe-max
open-swe-max-auto
pending
performance
pull-request
Mirrored from GitHub Pull Request
 question
requires design
sdk
server
stale
studio
tests
type checking
usage question
v0.5
v1.1
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langchain-ai/langgraph#638
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?