supporting prompt cache results show

lib/crewai/src/crewai/llms/providers/anthropic/completion.py

@@ -1580,10 +1580,12 @@ def _extract_anthropic_token_usage(
             usage = response.usage
             input_tokens = getattr(usage, "input_tokens", 0)
             output_tokens = getattr(usage, "output_tokens", 0)
+            cache_read_tokens = getattr(usage, "cache_read_input_tokens", 0) or 0
             return {
                 "input_tokens": input_tokens,
                 "output_tokens": output_tokens,
                 "total_tokens": input_tokens + output_tokens,
+                "cached_prompt_tokens": cache_read_tokens,
             }
         return {"total_tokens": 0}
 

lib/crewai/src/crewai/llms/providers/azure/completion.py

@@ -425,8 +425,9 @@ def _prepare_completion_params(
             "stream": self.stream,
         }
 
+        model_extras: dict[str, Any] = {}
         if self.stream:
-            params["model_extras"] = {"stream_options": {"include_usage": True}}
+            model_extras["stream_options"] = {"include_usage": True}
 
         if response_model and self.is_openai_model:
             model_description = generate_model_description(response_model)
@@ -464,6 +465,13 @@ def _prepare_completion_params(
             params["tools"] = self._convert_tools_for_interference(tools)
             params["tool_choice"] = "auto"
 
+        prompt_cache_key = self.additional_params.get("prompt_cache_key")
+        if prompt_cache_key:
+            model_extras["prompt_cache_key"] = prompt_cache_key
+
+        if model_extras:
+            params["model_extras"] = model_extras
+
         additional_params = self.additional_params
         additional_drop_params = additional_params.get("additional_drop_params")
         drop_params = additional_params.get("drop_params")
@@ -1063,10 +1071,15 @@ def _extract_azure_token_usage(response: ChatCompletions) -> dict[str, Any]:
         """Extract token usage from Azure response."""
         if hasattr(response, "usage") and response.usage:
             usage = response.usage
+            cached_tokens = 0
+            prompt_details = getattr(usage, "prompt_tokens_details", None)
+            if prompt_details:
+                cached_tokens = getattr(prompt_details, "cached_tokens", 0) or 0
             return {
                 "prompt_tokens": getattr(usage, "prompt_tokens", 0),
                 "completion_tokens": getattr(usage, "completion_tokens", 0),
                 "total_tokens": getattr(usage, "total_tokens", 0),
+                "cached_prompt_tokens": cached_tokens,
             }
         return {"total_tokens": 0}
 

lib/crewai/src/crewai/llms/providers/gemini/completion.py

@@ -1295,11 +1295,13 @@ def _extract_token_usage(response: GenerateContentResponse) -> dict[str, Any]:
         """Extract token usage from Gemini response."""
         if response.usage_metadata:
             usage = response.usage_metadata
+            cached_tokens = getattr(usage, "cached_content_token_count", 0) or 0
             return {
                 "prompt_token_count": getattr(usage, "prompt_token_count", 0),
                 "candidates_token_count": getattr(usage, "candidates_token_count", 0),
                 "total_token_count": getattr(usage, "total_token_count", 0),
                 "total_tokens": getattr(usage, "total_token_count", 0),
+                "cached_prompt_tokens": cached_tokens,
             }
         return {"total_tokens": 0}
 

lib/crewai/src/crewai/llms/providers/openai/completion.py

@@ -1094,11 +1094,7 @@ def _handle_streaming_responses(
                     if reasoning_items:
                         self._last_reasoning_items = reasoning_items
                 if event.response and event.response.usage:
-                    usage = {
-                        "prompt_tokens": event.response.usage.input_tokens,
-                        "completion_tokens": event.response.usage.output_tokens,
-                        "total_tokens": event.response.usage.total_tokens,
-                    }
+                    usage = self._extract_responses_token_usage(event.response)
                     self._track_token_usage_internal(usage)
 
         # If parse_tool_outputs is enabled, return structured result
@@ -1222,11 +1218,7 @@ async def _ahandle_streaming_responses(
                     if reasoning_items:
                         self._last_reasoning_items = reasoning_items
                 if event.response and event.response.usage:
-                    usage = {
-                        "prompt_tokens": event.response.usage.input_tokens,
-                        "completion_tokens": event.response.usage.output_tokens,
-                        "total_tokens": event.response.usage.total_tokens,
-                    }
+                    usage = self._extract_responses_token_usage(event.response)
                     self._track_token_usage_internal(usage)
 
         # If parse_tool_outputs is enabled, return structured result
@@ -1310,11 +1302,18 @@ def _extract_function_calls_from_response(
     def _extract_responses_token_usage(self, response: Response) -> dict[str, Any]:
         """Extract token usage from Responses API response."""
         if response.usage:
-            return {
+            result = {
                 "prompt_tokens": response.usage.input_tokens,
                 "completion_tokens": response.usage.output_tokens,
                 "total_tokens": response.usage.total_tokens,
             }
+            # Extract cached prompt tokens from input_tokens_details
+            input_details = getattr(response.usage, "input_tokens_details", None)
+            if input_details:
+                result["cached_prompt_tokens"] = (
+                    getattr(input_details, "cached_tokens", 0) or 0
+                )
+            return result
         return {"total_tokens": 0}
 
     def _extract_builtin_tool_outputs(self, response: Response) -> ResponsesAPIResult:
@@ -2264,11 +2263,18 @@ def _extract_openai_token_usage(
         """Extract token usage from OpenAI ChatCompletion or ChatCompletionChunk response."""
         if hasattr(response, "usage") and response.usage:
             usage = response.usage
-            return {
+            result = {
                 "prompt_tokens": getattr(usage, "prompt_tokens", 0),
                 "completion_tokens": getattr(usage, "completion_tokens", 0),
                 "total_tokens": getattr(usage, "total_tokens", 0),
             }
+            # Extract cached prompt tokens from prompt_tokens_details
+            prompt_details = getattr(usage, "prompt_tokens_details", None)
+            if prompt_details:
+                result["cached_prompt_tokens"] = (
+                    getattr(prompt_details, "cached_tokens", 0) or 0
+                )
+            return result
         return {"total_tokens": 0}
 
     def _format_messages(self, messages: str | list[LLMMessage]) -> list[LLMMessage]: