Feature: Show context/output limits in /models command

Feature Request

Problem

The /models command in the TUI currently only lists model names. There is no way to see the maximum context length or output token limit for each model without manually opening the config file or querying the provider API.

This makes it difficult to:

• Know which models support longer conversations
• Compare context windows when switching models
• Verify that config limits match the actual provider limits

Proposal

Enhance the /models command to display context and output token limits alongside each model name, e.g.:

litellm/glm-5.2-local    context: 300,000    output: 128,000
litellm/qwen3.6-plus     context: 1,000,000   output: 64,000
litellm/claude-sonnet-5  context: 1,000,000   output: 128,000

These values are already configured in opencode.json(c) under models.<name>.limit.context and models.<name>.limit.output, so the data is available locally without additional API calls.

Use Case

When working with long conversations or large codebases, users need to quickly identify which models have sufficient context capacity. Having this info inline in the /models picker would save time and avoid hitting context limits unexpectedly.