Model breakdown

Where to find it

The Models tab on Spend breaks the selected period down by model:

• •Calls per model
• •Cost per model
• •Average tokens per call
• •Share of spend

Finding opportunities

Look for expensive models handling simple tasks. Common patterns:

• •A flagship model for summarization - Claude Sonnet 4.6 lists at 40% less per token than Claude Opus 4.8, and Claude Haiku 4.5 at 80% less; whether quality holds is task-dependent, so check the advisor's confidence score
• •Large models for classification - fast-tier models are often sufficient for labeling and routing
• •One default model everywhere - teams that have not set per-task model selection

The Save view turns these patterns into suggestions, and the pre-send advisor prices a specific call on each alternative before you send it.

Weekly digest

On paid plans, account admins get a weekly digest email summarizing the week's spend by model category (premium, standard and efficient) and by provider, with coaching tips.