fix: context limit display uses stale model after model switch

[Ivaloz]

Description

After switching models mid-session, the context limit percentage in the sidebar continues to display based on the previous model's context window, not the currently selected model.

Example

1. Start session with model A (128k context)
2. Use model A for several messages, context shows 50% used (64k/128k)
3. Switch to model B (32k context)
4. Context still shows 50% used, but should show 200% (64k/32k) or at least use model B's limit

Root Cause

In the sidebar context component, the context limit is calculated using the model from the last assistant message, not the currently selected model:

\\ ypescript
const model = props.api.state.provider.find((item) => item.id === last.providerID)?.models[last.modelID]
\\

Proposed Fix

Use the currently selected model for context limit display when available:

\\ ypescript
const currentModel = local.model.current()
const model = currentModel
? props.api.state.provider.find((item) => item.id === currentModel.providerID)?.models[currentModel.modelID]
: props.api.state.provider.find((item) => item.id === last.providerID)?.models[last.modelID]
\\

MiMo Code Reference

This fix has been implemented in MiMo Code PR #1522: XiaomiMiMo/MiMo-Code#1522

Environment

• OpenCode version: 1.17.8
• OS: Windows 11
• Terminal: PowerShell

[tiffanychum]

I'd like to work on this. Confirmed the root cause in packages/tui/src/feature-plugins/sidebar/context.tsx: the context percentage is computed from the model of the last assistant message (last.providerID/last.modelID), so after a mid-session model switch it keeps using the previous model's context window. Fix: prefer the currently selected model via useLocal().model.current() and fall back to the last message's model when none is selected. PR incoming.