llama-index-llms-openai 0.6.1__tar.gz → 0.6.3__tar.gz

{llama_index_llms_openai-0.6.1 → llama_index_llms_openai-0.6.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: llama-index-llms-openai
-Version: 0.6.1
+Version: 0.6.3
 Summary: llama-index llms openai integration
 Author: llama-index
 License-Expression: MIT

{llama_index_llms_openai-0.6.1 → llama_index_llms_openai-0.6.3}/llama_index/llms/openai/base.py

@@ -1004,7 +1004,7 @@ class OpenAI(FunctionCallingLLM):
             del llm_kwargs["tool_choice"]
         return llm_kwargs
 
-    def _should_use_structure_outputs(self):
+    def _should_use_structure_outputs(self) -> bool:
         return (
             self.pydantic_program_mode == PydanticProgramMode.DEFAULT
             and is_json_schema_supported(self.model)

{llama_index_llms_openai-0.6.1 → llama_index_llms_openai-0.6.3}/llama_index/llms/openai/responses.py

@@ -450,7 +450,8 @@ class OpenAIResponses(FunctionCallingLLM):
 
         return stream_complete_fn(prompt, **kwargs)
 
-    def _parse_response_output(self, output: List[ResponseOutputItem]) -> ChatResponse:
+    @staticmethod
+    def _parse_response_output(output: List[ResponseOutputItem]) -> ChatResponse:
         message = ChatMessage(role=MessageRole.ASSISTANT, blocks=[])
         additional_kwargs = {"built_in_tool_calls": []}
         tool_calls = []
@@ -526,7 +527,7 @@ class OpenAIResponses(FunctionCallingLLM):
         if self.track_previous_responses:
             self._previous_response_id = response.id
 
-        chat_response = self._parse_response_output(response.output)
+        chat_response = OpenAIResponses._parse_response_output(response.output)
         chat_response.raw = response
         chat_response.additional_kwargs["usage"] = response.usage
         if hasattr(response.usage.output_tokens_details, "reasoning_tokens"):
@@ -590,7 +591,6 @@ class OpenAIResponses(FunctionCallingLLM):
         elif isinstance(event, ResponseTextDeltaEvent):
             # Text content is being added
             delta = event.delta
-            blocks.append(TextBlock(text=delta))
         elif isinstance(event, ResponseImageGenCallPartialImageEvent):
             # Partial image
             if event.partial_image_b64:
@@ -653,6 +653,8 @@ class OpenAIResponses(FunctionCallingLLM):
             # Response is complete
             if hasattr(event, "response") and hasattr(event.response, "usage"):
                 additional_kwargs["usage"] = event.response.usage
+            resp = OpenAIResponses._parse_response_output(event.response.output)
+            blocks = resp.message.blocks
 
         return (
             blocks,
@@ -782,7 +784,7 @@ class OpenAIResponses(FunctionCallingLLM):
         if self.track_previous_responses:
             self._previous_response_id = response.id
 
-        chat_response = self._parse_response_output(response.output)
+        chat_response = OpenAIResponses._parse_response_output(response.output)
         chat_response.raw = response
         chat_response.additional_kwargs["usage"] = response.usage
 

{llama_index_llms_openai-0.6.1 → llama_index_llms_openai-0.6.3}/llama_index/llms/openai/utils.py

@@ -62,6 +62,8 @@ O1_MODELS: Dict[str, int] = {
     "gpt-5-nano": 400000,
     "gpt-5-nano-2025-08-07": 400000,
     "gpt-5-chat-latest": 400000,
+    "gpt-5-pro": 400000,
+    "gpt-5-pro-2025-10-06": 400000,
 }
 
 O1_MODELS_WITHOUT_FUNCTION_CALLING = {

{llama_index_llms_openai-0.6.1 → llama_index_llms_openai-0.6.3}/pyproject.toml

@@ -27,7 +27,7 @@ dev = [
 
 [project]
 name = "llama-index-llms-openai"
-version = "0.6.1"
+version = "0.6.3"
 description = "llama-index llms openai integration"
 authors = [{name = "llama-index"}]
 requires-python = ">=3.9,<4.0"