Support for reasoning field in Chat Completion and maintenance roadmap for 1.1.x branch

Description

Currently, Spring AI (tested on version 1.1.2) appears to support the reasoning_content field (widely used by DeepSeek), but it does not seem to capture or map the reasoning field, which is frequently used by other providers or inference engines like vLLM (especially for Qwen 3.5 and other reasoning-capable models).

When using models that return the thought process under the reasoning key in the API response, this information is currently lost or inaccessible through the standard AssistantMessage or ChatResponse objects.

I am currently restricted to using Spring AI 1.1.x because my project is based on Spring Boot 3.5.x.

Questions

1. Feature Support: Are there plans to add support for the reasoning field to achieve parity with reasoning_content?
2. Backporting: Will this addition be available in the 1.1.x branch, or is it exclusive to the 2.x development line?
3. 1.1.x Maintenance Lifecycle: What is the maintenance plan for the 1.1.x branch? Since it aligns with Spring Boot 3.x, will it continue to receive feature updates for new AI standards (like reasoning support), or is it moving into a maintenance-only mode?

Context

With the rapid evolution of reasoning models, having a unified way to access the thought process (regardless of whether the key is reasoning or reasoning_content) is crucial for building transparent AI applications.

Thank you for your guidance and support!

Start by reviewing the Chat Completion response mapping and the AssistantMessage or ChatResponse objects mentioned in the issue, then compare how reasoning_content is currently handled. Confirm the expected behavior for a reasoning field and whether the change belongs in the 1.1.x branch; done means the reasoning value is accessible consistently and the branch policy is documented.