Skip to content

Automate browser based workflows with AI

Track OpenRouter LLM token counts and cost in get_call_stats (#6427)

Cindy Lid2e9caf5c5c2
diff --git a/skyvern/forge/sdk/api/llm/api_handler_factory.py b/skyvern/forge/sdk/api/llm/api_handler_factory.py
index acf0a8827..e0c6b8dc7 100644
--- a/skyvern/forge/sdk/api/llm/api_handler_factory.py
+++ b/skyvern/forge/sdk/api/llm/api_handler_factory.py
@@ -513,6 +513,26 @@ class LLMAPIHandlerFactory:
return input_tokens, output_tokens, reasoning_tokens, cached_tokens
+ @staticmethod
+ def _extract_reported_usage_cost(response: ModelResponse | CustomStreamWrapper) -> float | None:
+ """Return provider-reported cost from response.usage.cost when present."""
+ if not hasattr(response, "usage") or not response.usage:
+ return None
+
+ usage = response.usage
+ cost = getattr(usage, "cost", None)
+ if cost is None and isinstance(usage, dict):
+ cost = usage.get("cost")
+ if cost is None:
+ return None
+
+ try:
+ cost_value = float(cost)
+ except (TypeError, ValueError):
+ return None
+
+ return cost_value if cost_value >= 0 else None
+
@staticmethod
def _apply_thinking_budget_optimization(
parameters: dict[str, Any], new_budget: int, llm_config: LLMConfig | LLMRouterConfig, prompt_name: str
@@ -2875,8 +2895,6 @@ class LLMCaller:
self, response: ModelResponse | CustomStreamWrapper | AnthropicMessage | UITarsResponse
) -> LLMCallStats:
empty_call_stats = LLMCallStats()
- if self.original_llm_key.startswith("openrouter/"):
- return empty_call_stats
# Handle OPENAI_COMPATIBLE provider GitHub Copilot
if self.original_llm_key == "OPENAI_COMPATIBLE" and isinstance(response, (ModelResponse, CustomStreamWrapper)):
@@ -2917,14 +2935,30 @@ class LLMCaller:
reasoning_tokens=0,
)
elif isinstance(response, (ModelResponse, CustomStreamWrapper)):
- try:
- llm_cost = litellm.completion_cost(completion_response=response)
- except Exception as e:
- LOG.debug("Failed to calculate LLM cost", error=str(e), exc_info=True)
- llm_cost = 0
input_tokens, output_tokens, reasoning_tokens, cached_tokens = LLMAPIHandlerFactory._extract_token_counts(
response
)
+ llm_cost = 0.0
+ if self.original_llm_key.startswith("openrouter/"):
+ reported_cost = LLMAPIHandlerFactory._extract_reported_usage_cost(response)
+ if reported_cost is not None:
+ llm_cost = reported_cost
+ else:
+ LOG.warning(
+ "OpenRouter response missing usage.cost; cost will be reported as 0",
+ llm_key=self.original_llm_key,
+ response_id=getattr(response, "id", None),
+ )
+ else:
+ try:
+ llm_cost = litellm.completion_cost(completion_response=response)
+ except Exception as e:
+ LOG.debug(
+ "Failed to calculate LLM cost",
+ error=str(e),
+ exc_info=True,
+ )
+ llm_cost = 0
return LLMCallStats(
llm_cost=llm_cost,
input_tokens=input_tokens,