diff --git a/litellm/proxy/_types.py b/litellm/proxy/_types.py index e4b28001d..6196f18a2 100644 --- a/litellm/proxy/_types.py +++ b/litellm/proxy/_types.py @@ -151,6 +151,7 @@ class GenerateRequestBase(LiteLLMBase): rpm_limit: Optional[int] = None budget_duration: Optional[str] = None allowed_cache_controls: Optional[list] = [] + soft_budget: Optional[float] = None class GenerateKeyRequest(GenerateRequestBase): diff --git a/litellm/proxy/proxy_server.py b/litellm/proxy/proxy_server.py index dcd4283ba..482397b86 100644 --- a/litellm/proxy/proxy_server.py +++ b/litellm/proxy/proxy_server.py @@ -1995,6 +1995,9 @@ async def generate_key_helper_fn( except Exception as e: traceback.print_exc() raise HTTPException(status_code=status.HTTP_500_INTERNAL_SERVER_ERROR) + + # Add budget related info in key_data - this ensures it's returned + key_data["soft_budget"] = key_soft_budget return key_data