using langchain on python 3.13. This has been happening consistently for the past week.
my code is passing question to embedding for RAG retreival. The error reports via the google langchain code.
**GoogleGenerativeAIError(“Error embedding content: 500 INTERNAL. {‘error’: {‘code’: 500, ‘message’: ‘Internal error encountered.’, ‘status’: ‘INTERNAL’}}”)Traceback (most recent call last):
File “C:\Users\cdturner567\AppData\Local\Python\pythoncore-3.13-64\Lib\site-packages\langchain_google_genai\embeddings.py”, line 480, in embed_query
result = self.client.models.embed_content(
model=self.model,
contents=text,
config=config,
)**
langchain trace:
INFO - HTTP Request: POST (https)://generativelanguage.googleapis.com/v1beta/models/gemini-embedding-001:batchEmbedContents “HTTP/1.1 500 Internal Server Error”
DEBUG - receive_response_body.started request=<Request [b’POST’]>
DEBUG - receive_response_body.complete
DEBUG - response_closed.started
DEBUG - response_closed.complete
CRITICAL - Error processing chat request: Error embedding content: 500 INTERNAL. {‘error’: {‘code’: 500, ‘message’: ‘Internal error encountered.’, ‘status’: ‘INTERNAL’}}
2026-03-20-21:35:16 ChatServiceV1Gemini FATAL: Error processing chat request: Error embedding content: 500 INTERNAL. {‘error’: {‘code’: 500, ‘message’: ‘Internal error encountered.’, ‘status’: ‘INTERNAL’}}
using langsmith to trace the run, here is the Json passed:
{
“chat_history”: [
{
“additional_kwargs”: {},
“content”: “You are an expert Q&A system. Your task is to answer the user’s question using only the provided context.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “use a formal, and helpful tone.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “Instructions: Read the provided context carefully.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “If the answer is present in the context, provide a concise and direct answer.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “If the answer is not in the context, you MUST clearly state it.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “If there is not enough information in the context, you MUST clearly state it.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “If the question cannot be answered from the context, you MUST clearly state it.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “If you are unable to answer the question from the given context, you MUST clearly state it.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “Do not guess, fabricate information, or use your general knowledge. Do not use your internal knowledge.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “If you are describing Training in Power, use a positive tone.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “If there are multiple answers provide them in list format.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “If a person asks about classes, they are asking about events.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “If you need more context, please specify what additional information would help you provide a better answer.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “Assume the question is about Training in Power Academy unless otherwise stated.”,
“response_metadata”: {},
“type”: “system”
},
{
“additional_kwargs”: {},
“content”: “who is the president”,
“response_metadata”: {},
“type”: “human”
},
{
“additional_kwargs”: {},
“content”: “Sorry, there was an error processing your request. Please try again later.
“invalid_tool_calls”: ,
“response_metadata”: {},
“tool_calls”: ,
“type”: “ai”
},
{
“additional_kwargs”: {},
“content”: “who is the president”,
“response_metadata”: {},
“type”: “human”
},
{
“additional_kwargs”: {},
“content”: “Sorry, there was an error processing your request. Please try again later. “invalid_tool_calls”: ,
“response_metadata”: {},
“tool_calls”: ,
“type”: “ai”
},
{
“additional_kwargs”: {},
“content”: “who is the president”,
“response_metadata”: {},
“type”: “human”
}
],
“question”: “who is the president”
}
Let me know what other information you need. As I said I can repro consistently at the moment.