[ENHANCEMENT] Add Xhigh to openai compatible endpoints

[Soorma718]

Problem (one or two sentences)

Currently using azure to acces Gpt-5.2 which has xhigh setting for reasoning. Currently openai endpoints were updated to include xhigh, but openai compatible endpoints only have low,medium,and high. Can we please add xhigh to openai compatible endpoints.

Context (who is affected and when)

Using azure 5.2

Desired behavior (conceptual, not technical)

xhigh setting option

Constraints / preferences (optional)

No response

Request checklist

• I've searched existing Issues and Discussions for duplicates
• This describes a specific problem with clear context and impact

Roo Code Task Links (optional)

No response

Acceptance criteria (optional)

No response

Proposed approach (optional)

No response

Trade-offs / risks (optional)

No response

I've analyzed this issue and here's my proposed implementation plan:

Root Cause:
In OpenAICompatible.tsx, the ThinkingBudget component is passed supportsReasoningEffort: true (boolean). When ThinkingBudget receives a boolean true, it uses the base reasoningEfforts array from the types package, which is ["low", "medium", "high"] - not including "xhigh".

Proposed Solution:
Modify line 283 in OpenAICompatible.tsx to pass an explicit array including "xhigh":

// Change from:
supportsReasoningEffort: true,

// To:
supportsReasoningEffort: ["low", "medium", "high", "xhigh"],

Verification:

• The type schema already supports "xhigh" in the supportsReasoningEffort array (see model.ts)
• Translations for "xhigh" already exist in all 18 locale files (e.g., "Extra High" in English, "Sehr hoch" in German, etc.)
• The extended reasoning efforts type (reasoningEffortsExtended) includes "xhigh"

This is a minimal, focused change that will add the "Extra High" reasoning effort option to the OpenAI Compatible provider settings dropdown.

Please mention @roomote in a comment saying that you approve this plan, and I'll proceed with the implementation.

[mrubens]

@roomote go for it

Thanks for the approval! I'll proceed with the implementation now.

Opened PR #10061 with the implementation to add xhigh reasoning effort to OpenAI compatible endpoints. Feedback welcome!