Model usage overview
Model usage visuals show which engines you lean on most and how they contribute day to day. Use them to balance experimentation with reliable defaults.
Key metrics
Usage by model (pie)
- Share of usage: Percentage of total generations per model in the selected 7- or 30-day range
- Top models: Quickly spot the engines you use most
Daily contribution (stacked bars)
- Per-day mix: Daily counts for your top three models
- Spikes and dips: Identify days when a particular model dominated or dropped off
Recent generations
- Model context: Each recent item shows the model used, helping you connect outcomes to engine choice
- Cost visibility: Per-generation cost is visible without leaving the dashboard
Using model analytics
Identifying favorites
- Confirm which models dominate your usage.
- Spot underused models you might want to include in a multi-model batch for comparison.
Optimization opportunities
- Balance workload: If one model owns most of the pie, test alternatives on the same prompt.
- Validate results: Pair the dashboard view with session history to see which models produced the strongest outputs for a prompt family.
Best practices
Track over time: Compare 7-day and 30-day slices to see if your preferences are shifting.
Experiment: Run occasional multi-model batches when the pie chart shows a clear favorite to ensure it’s still your best option.
Balance usage: Keep multiple models in rotation so you have options when style or prompt requirements change.
