diff --git a/litellm/cost_calculator.py b/litellm/cost_calculator.py index 490f0288b00..71a93b757ce 100644 --- a/litellm/cost_calculator.py +++ b/litellm/cost_calculator.py @@ -36,6 +36,9 @@ from litellm.llms.anthropic.cost_calculation import ( from litellm.llms.azure.cost_calculation import ( cost_per_token as azure_openai_cost_per_token, ) +from litellm.llms.azure_ai.cost_calculator import ( + cost_per_token as azure_ai_cost_per_token, +) from litellm.llms.base_llm.search.transformation import SearchResponse from litellm.llms.bedrock.cost_calculation import ( cost_per_token as bedrock_cost_per_token, @@ -427,8 +430,8 @@ def cost_per_token( # noqa: PLR0915 return dashscope_cost_per_token(model=model, usage=usage_block) elif custom_llm_provider == "azure_ai": - return generic_cost_per_token( - model=model, usage=usage_block, custom_llm_provider=custom_llm_provider + return azure_ai_cost_per_token( + model=model, usage=usage_block, response_time_ms=response_time_ms ) else: model_info = _cached_get_model_info_helper( @@ -805,6 +808,7 @@ def _store_cost_breakdown_in_logging_obj( completion_tokens_cost_usd_dollar: float, cost_for_built_in_tools_cost_usd_dollar: float, total_cost_usd_dollar: float, + additional_costs: Optional[dict] = None, original_cost: Optional[float] = None, discount_percent: Optional[float] = None, discount_amount: Optional[float] = None, @@ -821,6 +825,7 @@ def _store_cost_breakdown_in_logging_obj( completion_tokens_cost_usd_dollar: Cost of completion tokens (includes reasoning if applicable) cost_for_built_in_tools_cost_usd_dollar: Cost of built-in tools total_cost_usd_dollar: Total cost of request + additional_costs: Free-form additional costs dict (e.g., {"azure_model_router_flat_cost": 0.00014}) original_cost: Cost before discount discount_percent: Discount percentage applied (0.05 = 5%) discount_amount: Discount amount in USD @@ -838,6 +843,7 @@ def _store_cost_breakdown_in_logging_obj( output_cost=completion_tokens_cost_usd_dollar, total_cost=total_cost_usd_dollar, cost_for_built_in_tools_cost_usd_dollar=cost_for_built_in_tools_cost_usd_dollar, + additional_costs=additional_costs, original_cost=original_cost, discount_percent=discount_percent, discount_amount=discount_amount, @@ -1335,6 +1341,21 @@ def completion_cost( # noqa: PLR0915 service_tier=service_tier, response=completion_response, ) + + # Get additional costs (e.g., Azure Model Router flat cost for azure_ai provider) + additional_costs: Optional[dict] = None + if custom_llm_provider == "azure_ai": + from litellm.llms.azure_ai.cost_calculator import ( + calculate_azure_model_router_flat_cost, + ) + azure_router_flat_cost = calculate_azure_model_router_flat_cost( + model=model, prompt_tokens=prompt_tokens + ) + if azure_router_flat_cost > 0: + additional_costs = { + "Azure Model Router Flat Cost": azure_router_flat_cost + } + _final_cost = ( prompt_tokens_cost_usd_dollar + completion_tokens_cost_usd_dollar ) @@ -1374,6 +1395,7 @@ def completion_cost( # noqa: PLR0915 completion_tokens_cost_usd_dollar=completion_tokens_cost_usd_dollar, cost_for_built_in_tools_cost_usd_dollar=cost_for_built_in_tools, total_cost_usd_dollar=_final_cost, + additional_costs=additional_costs, original_cost=original_cost, discount_percent=discount_percent, discount_amount=discount_amount, diff --git a/litellm/llms/azure_ai/cost_calculator.py b/litellm/llms/azure_ai/cost_calculator.py new file mode 100644 index 00000000000..b6258425a1f --- /dev/null +++ b/litellm/llms/azure_ai/cost_calculator.py @@ -0,0 +1,102 @@ +""" +Azure AI cost calculation helper. +Handles Azure AI Foundry Model Router flat cost and other Azure AI specific pricing. +""" + +from typing import Optional, Tuple + +from litellm._logging import verbose_logger +from litellm.litellm_core_utils.llm_cost_calc.utils import generic_cost_per_token +from litellm.types.utils import Usage +from litellm.utils import get_model_info + + +def _is_azure_model_router(model: str) -> bool: + """ + Check if the model is Azure AI Foundry Model Router. + + Detects patterns like: + - "azure-model-router" + - "model-router" + - "model_router/" + - "model-router/" + + Args: + model: The model name + + Returns: + bool: True if this is a model router model + """ + model_lower = model.lower() + return ( + "model-router" in model_lower + or "model_router" in model_lower + or model_lower == "azure-model-router" + ) + + +def calculate_azure_model_router_flat_cost(model: str, prompt_tokens: int) -> float: + """ + Calculate the flat cost for Azure AI Foundry Model Router. + + Args: + model: The model name (should be a model router model) + prompt_tokens: Number of prompt tokens + + Returns: + float: The flat cost in USD, or 0.0 if not applicable + """ + if not _is_azure_model_router(model): + return 0.0 + + # Get the model router pricing from model_prices_and_context_window.json + # Use "model_router" as the key (without actual model name suffix) + model_info = get_model_info(model="model_router", custom_llm_provider="azure_ai") + router_flat_cost_per_token = model_info.get("input_cost_per_token", 0) + + if router_flat_cost_per_token > 0: + return prompt_tokens * router_flat_cost_per_token + + return 0.0 + + +def cost_per_token( + model: str, usage: Usage, response_time_ms: Optional[float] = 0.0 +) -> Tuple[float, float]: + """ + Calculate the cost per token for Azure AI models. + + For Azure AI Foundry Model Router: + - Adds a flat cost of $0.14 per million input tokens (from model_prices_and_context_window.json) + - Plus the cost of the actual model used (handled by generic_cost_per_token) + + Args: + model: str, the model name without provider prefix + usage: LiteLLM Usage block + response_time_ms: Optional response time in milliseconds + + Returns: + Tuple[float, float] - prompt_cost_in_usd, completion_cost_in_usd + """ + # Calculate base cost using generic cost calculator + prompt_cost, completion_cost = generic_cost_per_token( + model=model, + usage=usage, + custom_llm_provider="azure_ai", + ) + + # Add flat cost for Azure Model Router + # The flat cost is defined in model_prices_and_context_window.json for azure_ai/azure-model-router + if _is_azure_model_router(model): + router_flat_cost = calculate_azure_model_router_flat_cost(model, usage.prompt_tokens) + + if router_flat_cost > 0: + verbose_logger.debug( + f"Azure AI Model Router flat cost: ${router_flat_cost:.6f} " + f"({usage.prompt_tokens} tokens × ${router_flat_cost / usage.prompt_tokens:.9f}/token)" + ) + + # Add flat cost to prompt cost + prompt_cost += router_flat_cost + + return prompt_cost, completion_cost