certain GenerateContentConfig params are incompatible with current client API

[punamateo]

I am trying to call the completions API with the following generation config, but I am getting a Pydantic error saying that the config params response_mime_type, system_instruction and response_schema are invalid, which is inconsistent with the documentation as of today and the code in the lib itself:

python-genai/google/genai/types.py

Line 3928 in 4193812

class GenerateContentConfig(_common.BaseModel):

Using API v1, but same error occurred using v1alpha.

To me, it's unclear whether the error is in the lib or in the API that is still updating, or even the model used, hence the need to submit the issue here.

My error:

Setting all components as failed. Reason: Gemini API request failed: 400 INVALID_ARGUMENT. {'error': {'code': 400, 'message': 'Invalid JSON payload received. Unknown name \"systemInstruction\": Cannot find field.\\nInvalid JSON payload received. Unknown name \"responseMimeType\" at \\'generation_config\\': Cannot find field.\\nInvalid JSON payload received. Unknown name \"responseSchema\" at \\'generation_config\\': Cannot find field.', 'status': 'INVALID_ARGUMENT', 'details': [{'@type': 'type.googleapis.com/google.rpc.BadRequest', 'fieldViolations': [{'description': 'Invalid JSON payload received. Unknown name \"systemInstruction\": Cannot find field.'}, {'field': 'generation_config', 'description': 'Invalid JSON payload received. Unknown name \"responseMimeType\" at \\'generation_config\\': Cannot find field.'}, {'field': 'generation_config', 'description': 'Invalid JSON payload received. Unknown name \"responseSchema\" at \\'generation_config\\': Cannot find field.'}]}]}}

Environment details

• Programming language: Python
• OS: Linux
• Language runtime version:3.11
• Package version: 1.32.0

Steps to reproduce

1. Initialize response_schema and google-genai client

        self.client = genai.Client(api_key=self.api_key,http_options=HttpOptions(api_version='v1')) if self.api_key else None # Using API version v1

            response_schema = {
                "type": "object",
                "properties": {
                    "screens": {
                        "type": "array",
                        "items": {
                            "type": "object",
                            "properties": {
                                "screen_name": {"type": "string"},
                                "sub_prompt": {"type": "string"}
                            },
                            "required": ["screen_name", "sub_prompt"]
                        }
                    }
                },
                "required": ["screens"]
            }

2. Set GenerationConfig

         generation_config = GenerateContentConfig(
                temperature=1,
                max_output_tokens=10000,
                response_mime_type="application/json",
                system_instruction=formatted_messages["system_instruction"],
                safety_settings=safety_settings,
                response_schema=response_schema
            )

3. Call the generate_content function with gemini 2.5-pro

            response = self.client.models.generate_content(
                model="models/gemini-2.5-pro",
                contents=formatted_messages["contents"],
                config=generation_config
            )

[janasangeetha]

Hi @punamateo
Thanks for reaching out!
I believe this is same as #1205

[github-actions]

This issue has been marked as stale because it has been open for 7 days with no activity. It will be closed in 2 days if no further activity occurs.

[github-actions]

This issue was closed because it has been inactive for 9 days. Please post a new issue if you need further assistance. Thanks!