## Added

- Added the official Codex `gpt-6.1-sol` model template with its official prompts and capabilities. The default reasoning effort is `low`, with six supported levels: `low / medium / high / xhigh / max / ultra`. The default context window is 272000 tokens, with a maximum of 872000 tokens.

## Improved

- Updated the embedded Codex model catalog revision to prevent older cached catalogs from overwriting the new model template.

## Upcoming

- We plan to redesign the UI in v0.3.13 to further improve the user experience. Please share any suggestions through GitHub Issues.