See your AI token utilization by model

Account → Usage shows a Token utilization card for the Claude model your org has selected for Wally. You'll see input, output, cache-read, and total tokens used this month, a breakdown bar, your turn count, and a cache-hit note (prompt caching lowers your cost).

The card is labeled with your active model — Turtini-hosted, or your bring-your-own preset (Bedrock, Vertex, etc.) with its model id and region. It sits alongside the pay-as-you-go card so you can see both the tokens consumed and what they cost (at cost + 15%).