Unable to test openai/gpt-oss-120b via vllm

[psydok]

Describe the bug
KeyError: 'content'

Expected behavior
Can it skip chunks where content is not specified?

Environment
Include all relevant environment information:

1. OS [e.g. Ubuntu 20.04]: -
2. Python version [e.g. 3.12.2]: 3.10

To Reproduce
Exact steps to reproduce the behavior:

1. Run vllm on H100

vastai create instance <OFFER_ID> --image vllm/vllm-openai:gptoss --env '-p 8000:8000 --ipc=host --gpus all' --disk 8 --args --model "openai/gpt-oss-120b" --tensor-parallel-size 1

2. Run guidellm

export GUIDELLM__PREFERRED_ROUTE="chat_completions" && export GUIDELLM__OPENAI__MAX_OUTPUT_TOKENS=512 && export GUIDELLM__MAX_CONCURRENCY=233 && export GUIDELLM__REQUEST_TIMEOUT=300 && export GUIDELLM__PREFERRED_PROMPT_TOKENS_SOURCE=local && export GUIDELLM__PREFERRED_OUTPUT_TOKENS_SOURCE=local && guidellm benchmark --target http://89.25.97.3:11509 --rate-type sweep --rate 5 --model openai/gpt-oss-120b --processor openai/gpt-oss-120b --random-seed 2025 --max-requests=100 --data "prompt_tokens=4096,output_tokens=512" --backend-args '{"extra_body":{"chat_template_kwargs":{"enable_thinking":false}}}' --output-path "data/benchmarks.json"

Errors

Creating backend...
2025-08-15T09:46:20.216941+0000 | chat_completions | ERROR - OpenAIHTTPBackend request with headers: {'Content-Type': 'application/json'} and params: {} and payload: {'chat_template_kwargs': {'enable_thinking': False}, 'messages': [{'role': 'user', 'content': 'Test connection'}], 'model': 'openai/gpt-oss-120b', 'stream': True, 'stream_options': {'include_usage': True}, 'max_tokens': 1, 'max_completion_tokens': 1, 'stop': None, 'ignore_eos': True} failed: 'content'
Traceback (most recent call last):
  File "/home/user/.local/bin/guidellm", line 7, in <module>
    sys.exit(cli())
  File "/home/user/.local/lib/python3.10/site-packages/click/core.py", line 1161, in __call__
    return self.main(*args, **kwargs)
  File "/home/user/.local/lib/python3.10/site-packages/click/core.py", line 1082, in main
    rv = self.invoke(ctx)
  File "/home/user/.local/lib/python3.10/site-packages/click/core.py", line 1697, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/home/user/.local/lib/python3.10/site-packages/click/core.py", line 1697, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/home/user/.local/lib/python3.10/site-packages/click/core.py", line 1443, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/home/user/.local/lib/python3.10/site-packages/click/core.py", line 788, in invoke
    return __callback(*args, **kwargs)
  File "/home/user/.local/lib/python3.10/site-packages/guidellm/__main__.py", line 314, in run
    asyncio.run(
  File "/usr/local/lib/python3.10/asyncio/runners.py", line 44, in run
    return loop.run_until_complete(main)
  File "/usr/local/lib/python3.10/asyncio/base_events.py", line 649, in run_until_complete
    return future.result()
  File "/home/user/.local/lib/python3.10/site-packages/guidellm/benchmark/entrypoints.py", line 29, in benchmark_with_scenario
    return await benchmark_generative_text(**vars(scenario), **kwargs)
  File "/home/user/.local/lib/python3.10/site-packages/guidellm/benchmark/entrypoints.py", line 71, in benchmark_generative_text
    await backend.validate()
  File "/home/user/.local/lib/python3.10/site-packages/guidellm/backend/backend.py", line 138, in validate
    async for _ in self.chat_completions(  # type: ignore[attr-defined]
  File "/home/user/.local/lib/python3.10/site-packages/guidellm/backend/openai.py", line 393, in chat_completions
    raise ex
  File "/home/user/.local/lib/python3.10/site-packages/guidellm/backend/openai.py", line 374, in chat_completions
    async for resp in self._iterative_completions_request(
  File "/home/user/.local/lib/python3.10/site-packages/guidellm/backend/openai.py", line 623, in _iterative_completions_request
    if delta := self._extract_completions_delta_content(type_, data):
  File "/home/user/.local/lib/python3.10/site-packages/guidellm/backend/openai.py", line 691, in _extract_completions_delta_content
    return data["choices"][0]["delta"]["content"]
KeyError: 'content'

Additional context
Add any other context about the problem here. Also include any relevant files.

[sjmonson]

Revisit after #251