fix(ai): parse OpenAI-compatible reasoning deltas (v2)

[bbartels]

Issue for this PR

Closes #37553
Fixes #37553

Type of change

• Bug fix
• New feature
• Refactor / code improvement
• Documentation

What does this PR do?

Some OpenAI-compatible providers return streamed reasoning in delta.reasoning instead of delta.reasoning_content.

This adds reasoning to the stream schema and parses it as a fallback when reasoning_content is absent. This matches the behavior of the AI SDK's OpenAI-compatible provider while preserving reasoning_content precedence.
This is a regression from v1 behaviour, v1 supports reasoning for vllm compatibility.

How did you verify your code works?

Added a regression test using an SSE stream containing delta.reasoning, followed by normal response text.

Ran from packages/ai:

bun test test/provider/openai-chat.test.ts

Screenshots / recordings
Not applicable; this is a protocol parsing fix.
Checklist

• I have tested my changes locally
• I have not included unrelated changes in this PR

[github-actions]

This PR doesn't fully meet our contributing guidelines and PR template.

What needs to be fixed:

• PR description is missing required template sections. Please use the PR template.

Please edit this PR description to address the above within 2 hours, or it will be automatically closed.

If you believe this was flagged incorrectly, please let a maintainer know.

[github-actions]

Thanks for your contribution!

This PR doesn't have a linked issue. All PRs must reference an existing issue.

Please:

1. Open an issue describing the bug/feature (if one doesn't exist)
2. Add Fixes #<number> or Closes #<number> to this PR description

See CONTRIBUTING.md for details.

[github-actions]

The following comment was made by an LLM, it may be inaccurate:

Potential Related PRs

I found several PRs related to OpenAI reasoning parsing that may be relevant:

1. PR fix: accept Ollama reasoning field in OpenAI Chat deltas #36068 - fix: accept Ollama reasoning field in OpenAI Chat deltas

   • Directly related: handles the reasoning field from Ollama in OpenAI Chat streams (similar fallback pattern)

2. PR fix(llm): accept reasoning field in OpenAI-compatible streams #35284 - fix(llm): accept 'reasoning' field in OpenAI-compatible streams

   • Directly related: also addresses accepting the reasoning field as a fallback in OpenAI-compatible streams

3. PR fix(llm): serialize overlapping OpenAI reasoning #37022 - fix(llm): serialize overlapping OpenAI reasoning

   • Related to reasoning serialization in OpenAI compatibility

These PRs (particularly #35284 and #36068) address very similar issues with parsing the reasoning field from OpenAI-compatible providers. You may want to check if this PR duplicates or should be combined with those fixes.

[bbartels]

Duplicate of: https://github.com/anomalyco/opencode/pull/35284/changes