O1-pro API not Calling Functions

I am using the same scaffolding using the responses API to have gpt-4o and o3-mini perform function calling (which all works fine), but o1-pro does not seem to actually call functions. It seems to have gotten all the information from the tool definitions and describes the function calls in its response, but is only providing a ResponseOutputText object in the output rather than a ResponseFunctionToolCall. See below for snippet of output (from response.output_text) indicates it is understanding the tool call definition.

…other response text above…

Below are function calls to execute the queries:

  1. Compare actual vs. expected sales across 2021–2023:

• Explanation: Summarize actual vs. expected sales totals by year to identify any significant deviations.

{
  "name": "query_data",
  "arguments": {
    "sql_query": "SELECT
  year,
  SUM(actual_sales) AS actual_sales,
  SUM(expected_sales) AS expected_sales,
  CASE 
      WHEN SUM(expected_sales) = 0 THEN NULL 
      ELSE SUM(actual_sales) / SUM(expected_sales) 
  END AS sales_ae_ratio
FROM sales_performance
WHERE year IN (2021, 2022, 2023)
GROUP BY year
ORDER BY year;",
    "title": "Compare Sales Performance A/E for 2021-2023"
  }
}

…more response text and “function calls” below…

This describes the proper JSON for the function call as defined in the tools definition (and only defined there, not in the developer message), but is not actually coming back as a function call for some reason. Is this a known issue?