[ENHANCEMENT] Enable model selection in chat for OpenAI compatible mode

[Ardakilic]

Problem (one or two sentences)

Heya,

I'd appreciate if model selection for OpenAI Compatible models is available on chat screen.

Context (who is affected and when)

Everyone as far as I can tell.

I tried with openrouter and Kilo gateway (as only two other openai compatible api gateways I have), both have this issue.

As you can see we can already select models on this page:

But not on chat screen.

Desired behavior (conceptual, not technical)

Chat screen to have model selection dropdown, just like settings page.

Constraints / preferences (optional)

Apologies if this is already opened, I looked for but couldn't find a similar issue.

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

[roomote-v0]

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

Problem: The OpenAI Compatible provider (and other non-static-model providers like Ollama, LM Studio) allows model selection on the settings page via ModelPicker, but the chat screen only has the ApiConfigSelector (which switches between API configuration profiles, not individual models). Users must navigate to settings to change their model.

Proposed approach:

1. Add a model picker to the chat input bar for providers that support dynamic model lists. This would appear alongside the existing ModeSelector and ApiConfigSelector in the bottom toolbar of ChatTextArea.tsx.

2. Extend useSelectedModel hook to also expose the available models list for the current provider, so the chat-level model picker can display options without re-fetching.

3. Create a lightweight ChatModelPicker component that:

   • Shows the currently selected model ID as a compact button (similar to how ApiConfigSelector works)
   • On click, opens a popover with the available models list
   • For OpenAI Compatible (openai provider): fetches models from the /models endpoint (reusing the existing openAiModels message mechanism)
   • For Ollama/LM Studio: uses the already-fetched dynamic model lists from useOllamaModels/useLmStudioModels
   • For static-model providers (Anthropic, Bedrock, etc.): shows the static model list from @roo-code/types
   • Allows typing a custom model ID (as the settings ModelPicker already does)

4. Wire the selection through setApiConfigurationField to update openAiModelId (or the appropriate model ID key per provider), so the change takes effect immediately for the next message.

5. Add tests for the new ChatModelPicker component.

Key files involved:

• webview-ui/src/components/chat/ChatTextArea.tsx - add the model picker to the toolbar
• New file: webview-ui/src/components/chat/ChatModelPicker.tsx
• webview-ui/src/components/ui/hooks/useSelectedModel.ts - potentially expose available models
• webview-ui/src/components/settings/ModelPicker.tsx - reference/reuse logic

This approach keeps the existing settings-page model picker intact and adds a convenient shortcut on the chat screen. It would work for all providers, not just OpenAI Compatible.

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