fix: response_format is ignored in Azure OpenAI models with API version >= 2025

[shionhonda]

Title

This PR fixes a bug in the function _is_response_format_supported_api_version.
Due to this bug, we are not passing the argument response_format in API version >= 2025.

litellm/litellm/llms/azure/chat/gpt_transformation.py

Lines 117 to 124 in 3875df6

	def _is_response_format_supported_api_version(
	self, api_version_year: str, api_version_month: str
	) -> bool:
	"""
	- check if api_version is supported for response_format
	"""
	is_supported = int(api_version_year) <= 2024 and int(api_version_month) >= 8

I believe it should evaluate true if it is >= "2024-08"

Relevant issues

• Azure api version check - fix str compare to convert to int #8438

Pre-Submission checklist

Please complete all items before asking a LiteLLM maintainer to review your PR

I followed the guideline but couldn't make the existing tests pass on the fresh main branch (Failed: Error occurred: litellm.APIError: AzureException APIError - argument of type 'NoneType' is not iterable)

Type

🐛 Bug Fix

Changes

• _is_response_format_supported_api_version should be true when the API version >= 2025
• Added test cases

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
litellm	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Mar 14, 2025 1:22pm

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[shionhonda]

Bug fixed in #9705