This update makes it easier to evaluate and choose models at a glance. You can now see benchmark performance details in the model picker, alongside a few UI touches that add clarity without getting in the way.
New in This Release
Model benchmarks are now integrated and kept up to date, so benchmark data is available when you browse models. This helps you compare options more confidently using consistent performance signals.
The model selector now includes a new info panel with benchmark performance display. It also adds clearer capability indicators, including a tool use badge, so you can quickly understand what a model supports before selecting it.
Polished UI Details
Forked chats are now easier to spot in the sidebar, thanks to a fork icon shown next to chats that were created from another conversation.
Tooltips gained more flexible positioning controls, which helps keep hover details readable and better placed. Thin scrollbars were also added in the UI for a cleaner, more compact browsing experience.