Repository navigation
Incomplete responses are not handled by parse_response when using Structured Output #2486
Description
Confirm this is an issue with the Python library and not an underlying OpenAI API
- This is an issue with the Python library
Describe the bug
I'm using Structured Outputs and have a max token length of 2000 for my finetuned model. When I generate a response, response.output has a length of 8 and the last output is incomplete. When response.output contains an incomplete output, parse_response doesn't handle this and instead raises a ValidationError.
It seems like there should either be a way to constrain the length of response.output or ignore incomplete outputs in parsing, for example:
if output.type == "message" and output.status != "incomplete":
To Reproduce
To repro, ensure that response.output contains an incomplete output and then call parse_response with a structured output.
raw_response.output[-1] = ResponseOutputMessage(id='xxx', content=[ResponseOutputText(annotations=[], text='{"a": {"b": 5.0, , "c": {"d": 4.0,', type='output_text', logprobs=[])], role='assistant', status='incomplete', type='message')
parse_response(
text_format=ExpectedResponse,
response=raw_response,
)
Code snippets
client = OpenAI()
resp = client.responses.parse(
model=model,
instructions=prompt,
input=input_text.to_list(),
temperature=temperature,
max_output_tokens=max_tokens
text_format=ExpectedResponse,
)
yields
File "/.venv/lib/python3.13/site-packages/openai/_base_client.py", line 1124, in _process_response
return api_response.parse()
~~~~~~~~~~~~~~~~~~^^
File "/venv/lib/python3.13/site-packages/openai/_response.py", line 325, in parse
parsed = self._options.post_parser(parsed)
File "/.venv/lib/python3.13/site-packages/openai/resources/responses/responses.py", line 936, in parser
return parse_response(
input_tools=tools,
text_format=text_format,
response=raw_response,
)
File "/.venv/lib/python3.13/site-packages/openai/lib/_parsing/_responses.py", line 75, in parse_response
"parsed": parse_text(item.text, text_format=text_format),
~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/.venv/lib/python3.13/site-packages/openai/lib/_parsing/_responses.py", line 137, in parse_text
return cast(TextFormatT, model_parse_json(text_format, text))
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^
File "/.venv/lib/python3.13/site-packages/openai/_compat.py", line 169, in model_parse_json
return model.model_validate_json(data)
~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^
File "/.venv/lib/python3.13/site-packages/pydantic/main.py", line 746, in model_validate_json
return cls.__pydantic_validator__.validate_json(
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^
json_data, strict=strict, context=context, by_alias=by_alias, by_name=by_name
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
pydantic_core._pydantic_core.ValidationError: 1 validation error for ExpectedResponse
Invalid JSON: EOF while parsing an object at line 1 column 234 [type=json_invalid, input_value='{"example_key":...n": {"clarity": 6', input_type=str]
For further information visit https://errors.pydantic.dev/2.11/v/json_invalidOS
macOS
Python version
python v3.13
Library version
openai v1.86.0
Activity
fix: handle incomplete JSON in parse_response for Structured Output (o…
nightcityblade commentedon May 20, 2026
Hi, I'd like to work on this. I'll submit a small PR shortly.
marcuswood-oai commentedon Sep 21, 2026
Tracking the streaming case from #3263 here alongside the non-streaming report. The fix needs to cover:
responses.parse(): preserve incomplete output items and their status without trying to parse their truncated structured text, while still parsing completed siblings.responses.stream(): an earlyresponse.output_text.doneparsing failure must not hide the terminalresponse.incompleteevent or its reason.get_final_response(): preserve the terminal incomplete response, including its content, status, and incomplete details, for both sync and async callers.Validation of completed structured output should remain strict. This issue stays open until both the streaming and non-streaming cases are addressed.