feat: /v1/responses qwen3 max && perplexity

[seefs001]

#2526
only qwen3-max-2026-01-23

Summary by CodeRabbit

• New Features

  • Added support for configuring "thinking" behavior and preset options in API requests.

• Improvements

  • Enhanced routing for response-mode requests to the appropriate responses endpoint.
  • Made response-mode request handling functional to improve end-to-end request flow and reliability.

[coderabbitai]

Walkthrough

Added two new JSON fields (EnableThinking, Preset) to the OpenAI responses request DTO; implemented RelayModeResponses URL routing and made ConvertOpenAIResponsesRequest a no-op passthrough in the Ali and Perplexity channel adaptors.

Changes

Cohort / File(s)	Summary
DTO Enhancement dto/openai_request.go	Added EnableThinking json.RawMessage and Preset json.RawMessage fields to OpenAIResponsesRequest (json tags enable_thinking and preset).
Ali Adapter relay/channel/ali/adaptor.go	Added RelayModeResponses branch to GetRequestURL to route to /api/v2/apps/protocols/compatible-mode/v1/responses; replaced not-implemented ConvertOpenAIResponsesRequest with passthrough returning the incoming request.
Perplexity Adapter relay/channel/perplexity/adaptor.go	Imported relay constants; added RelayModeResponses handling in GetRequestURL to route to /v1/responses; implemented ConvertOpenAIResponsesRequest to return the provided request unchanged.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• 修复Azure渠道对responses API的兼容性支持 - 为Azure渠道添加对responses API的特殊处理 - 兼容微软新… #1205: Adds special-case handling for RelayModeResponses in channel adaptor GetRequestURL implementations.
• feat: pplx channel #1998: Overlaps with Perplexity adaptor routing and request conversion changes and adds request fields in DTOs.
• #1570: Related changes to propagate thinking payloads across request/response pipelines (field propagation).

Poem

🐰 I nibble at fields both new and neat,
EnableThinking and Preset tucked in sweet,
Adapters hop paths with Responses in mind,
Requests pass through, no placeholder to find —
A tiny code-hop, efficient and fleet.

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Title check	✅ Passed	The title 'feat: /v1/responses qwen3 max && perplexity' accurately reflects the main changes: adding /v1/responses endpoint support for both Qwen3 Max and Perplexity models.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}