remove specific model checks we don't need

[justschen]

added in as a defensive thing but we don't really need and shouldn't do this

[Copilot]

Pull Request Overview

This pull request removes model-specific logic for handling bolded content in chat thinking parts. The change makes the bold stripping behavior universal rather than conditional on the model being "codex".

Key Changes

• Removed unused import isResponseVM from chatViewModel
• Removed the modelId private field and its initialization logic
• Made bold marker stripping (**...**) apply to all models instead of only "codex" models

[Copilot]

The bold stripping logic is too simplistic and may incorrectly strip valid markdown bold markers. If the content is 'Bold text more', this would incorrectly strip the outer markers leaving 'Bold** text **more' which is malformed markdown. Consider checking that the content doesn't contain any other bold markers between the start and end, or use a more robust regex-based approach to detect if the entire content is a single bold block.

Suggested change

	if (cleanedContent.startsWith('**') && cleanedContent.endsWith('**')) {
	contentToRender = cleanedContent.slice(2, -2);
	const boldBlockMatch = cleanedContent.match(/^\*\*(.+)\*\*$/s);
	if (boldBlockMatch) {
	contentToRender = boldBlockMatch[1];