[Bugfix] fix missing 'finish_reason': null in streaming chat

[chaunceyjiang]

FIX #19650

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[gemini-code-assist]

Summary of Changes

Hello @chaunceyjiang, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request addresses a bug in the streaming chat completion endpoint where the finish_reason field was missing from intermediate chunks. By changing how the response model is serialized, the PR ensures that this field is always present, typically as null, until the final chunk, improving compliance with the expected API output format.

Highlights

• Bugfix: Fixes an issue where the finish_reason field was incorrectly omitted from streaming chat completion chunks before the final chunk, instead of being included with a null value.
• API Compliance: Ensures that the streaming chat completion response format aligns more closely with the OpenAI API specification by always including the finish_reason field in each chunk.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in issue comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist is currently in preview and may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments to provide feedback.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configureGemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[chaunceyjiang]

Test

data: {"id":"chatcmpl-a0e8d16d47a64de58a0a3083e18d432a","object":"chat.completion.chunk","created":1749994809,"model":"/home/jovyan/qwen3-8b","choices":[{"index":0,"delta":{"content":"가"},"logprobs":null,"finish_reason":null}]}

data: {"id":"chatcmpl-a0e8d16d47a64de58a0a3083e18d432a","object":"chat.completion.chunk","created":1749994809,"model":"/home/jovyan/qwen3-8b","choices":[{"index":0,"delta":{"content":"요"},"logprobs":null,"finish_reason":null}]}

data: {"id":"chatcmpl-a0e8d16d47a64de58a0a3083e18d432a","object":"chat.completion.chunk","created":1749994809,"model":"/home/jovyan/qwen3-8b","choices":[{"index":0,"delta":{"content":"?"},"logprobs":null,"finish_reason":null}]}

data: {"id":"chatcmpl-a0e8d16d47a64de58a0a3083e18d432a","object":"chat.completion.chunk","created":1749994809,"model":"/home/jovyan/qwen3-8b","choices":[{"index":0,"delta":{"content":""},"logprobs":null,"finish_reason":"stop","stop_reason":null}]}

[gemini-code-assist]

Code Review

This pull request addresses an issue where finish_reason: null might be missing in streaming chat responses. The change modifies the Pydantic serialization behavior in vllm/entrypoints/openai/serving_chat.py by switching from exclude_none=True to exclude_unset=True when calling model_dump_json for intermediate stream chunks.

This change is correct for ensuring that fields explicitly set to None (like finish_reason = None for ongoing choices) are included in the JSON output with a null value, rather than being omitted. This aligns with typical API behavior where null is a meaningful state for optional fields.

The modification at line 876 primarily affects intermediate stream chunks, including those that provide continuous usage statistics. While the linked issue #19650 specifically mentions the "final chunk with usage info" (serialized at line 899, which already uses exclude_unset=True), this PR correctly addresses the finish_reason: null behavior for other streamed chunks, contributing to overall API consistency.

The change is minimal, targeted, and appears to be a solid fix.

[chaunceyjiang]

Test1

curl -XPOST http://localhost:8080/v1/chat/completions -H 'Content-Type: application/json' -d '{"model":"/home/jovyan/qwen3-8b","messages":[{"role":"user","content":"안녕?"}],"stream":true}'

data: {"id":"chatcmpl-a0e8d16d47a64de58a0a3083e18d432a","object":"chat.completion.chunk","created":1749994809,"model":"/home/jovyan/qwen3-8b","choices":[{"index":0,"delta":{"content":"가"},"logprobs":null,"finish_reason":null}]}

data: {"id":"chatcmpl-a0e8d16d47a64de58a0a3083e18d432a","object":"chat.completion.chunk","created":1749994809,"model":"/home/jovyan/qwen3-8b","choices":[{"index":0,"delta":{"content":"요"},"logprobs":null,"finish_reason":null}]}

data: {"id":"chatcmpl-a0e8d16d47a64de58a0a3083e18d432a","object":"chat.completion.chunk","created":1749994809,"model":"/home/jovyan/qwen3-8b","choices":[{"index":0,"delta":{"content":"?"},"logprobs":null,"finish_reason":null}]}

data: {"id":"chatcmpl-a0e8d16d47a64de58a0a3083e18d432a","object":"chat.completion.chunk","created":1749994809,"model":"/home/jovyan/qwen3-8b","choices":[{"index":0,"delta":{"content":""},"logprobs":null,"finish_reason":"stop","stop_reason":null}]}

Test2:

curl --location --request POST 'http://127.0.0.1:8000/v1/chat/completions' \
  --header 'Authorization: Bearer xxx' \
  --header 'Content-Type: application/json' \
  --data-raw '{
    "model": "",
    "messages": [{"role": "user", "content": "What is the weather like in Boston today?"}],
    "tools": [{
      "type": "function",
      "function": {
        "name": "get_current_weather",
        "description": "Get the current weather in a given location",
        "parameters": {
          "type": "object",
          "properties": {
            "location": {"type": "string", "description": "The city and state, e.g. San Francisco, CA"},
            "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]}
          },
          "required": ["location"]
        }
      }
    }],
    "tool_choice": {"type": "function", "function": {"name": "get_current_weather"}},
    "stream": true
  }'

data: {"id":"chatcmpl-2b45106e5bef4f1eabc563b3da1f2649","object":"chat.completion.chunk","created":1750067911,"model":"Qwen/Qwen3-8B","choices":[{"index":0,"delta":{"tool_calls":[{"id":"chatcmpl-tool-040af537f3e14bd6986646facde02ce3","type":"function","index":0,"function":{"name":"get_current_weather","arguments":"{\n\n"}}]},"logprobs":null,"finish_reason":null}]}

data: {"id":"chatcmpl-2b45106e5bef4f1eabc563b3da1f2649","object":"chat.completion.chunk","created":1750067911,"model":"Qwen/Qwen3-8B","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\""}}]},"logprobs":null,"finish_reason":null}]}

data: {"id":"chatcmpl-2b45106e5bef4f1eabc563b3da1f2649","object":"chat.completion.chunk","created":1750067911,"model":"Qwen/Qwen3-8B","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"location"}}]},"logprobs":null,"finish_reason":null}]}

You can see that "finish_reason": null is consistent with OpenAI's behavior, and the tool_calls structure is also aligned with OpenAI.

OpenAI standard

data: {"id":"chatcmpl-BKLwktzPcPGbK0ha9zurbznPoRDYD","object":"chat.completion.chunk","created":1744190622,"model":"gpt-4o-mini-2024-07-18","choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_qj5ZjQ2f60WZVrKUkG1LgWxK","type":"function","function":{"name":"get_current_weather","arguments":""}}],"logprobs":null,"finish_reason":null}]},"usage":null}

[chaunceyjiang]

/cc @maxdebayser @cjackal @aarnphm @DarkLight1337 PTAL.

[DarkLight1337]

Thanks for fixing!