ModelHTTPError: status_code: 400, model_name: ai_model, body: {'message': '1 validation error for list[function-wrap[log_extra_fields()]]\n Invalid JSON: EOF while parsing a string at line 6 column 952 [type=json_invalid, input_value='[\n\n{\n "name": "fin...cosmetic). Although the', input_type=str]\n For further information visit https://errors.pydantic.dev/2.12/v/json_invalid', 'type': 'BadRequestError', 'param': None, 'code': 400}

Fix Plan

To resolve the issue, we need to increase the max_len parameter to a higher value to accommodate the thinking content generated by the model. Here are the steps:

• Increase the max_len parameter to a higher value, e.g., max_len=8000 or more, depending on the specific use case.
• Alternatively, consider implementing a more dynamic approach to handle the thinking content, such as:

if tool_choice == "required" and thinking_enabled:
    max_len = 8000  # or a higher value

• If using a configuration file, update the max_len value accordingly.

Verification

To verify that the fix worked, run the model with the increased max_len value and check for the following:

• The model should be able to generate the thinking content without exceeding the max_len limit.
• The stop_reason should not be length.
• The tool should be called successfully, and the output should be in the expected format.

Extra Tips

• Consider adding a check to handle cases where the thinking content exceeds the max_len limit, to prevent similar issues in the future.
• Review the model's configuration and adjust the max_len value as needed to balance between performance and output quality.