# 17. Your model-choice and usage card

The [model and allowance walkthrough](https://cowanworks.com/aicourse/learn/course-1/lesson-20) uses ordinary writing and a constrained workshop comparison.

| Task | Starting capability | Success check |
| --- | --- | --- |
| Shorten a supplied message | Default chat | Preserve the event and reply deadline |
| Compare workshop options | Reasoning-capable mode if available | A costs $120 and needs 120 minutes; B costs $210 and needs 75; neither fits $180 and 100 minutes |
| Find a current public price | Search with direct sources | Open the actual current offer and check terms |

For your own task, record: product and workspace; model/mode if visible; input source; result required; how you checked it; waiting time; and any displayed usage notice. One trial does not establish universal model quality.

```
Before recommending a different model or mode, identify the
capability this task needs, the information still missing and
how I can judge the result. Do not invent plan features or prices.
```

Inspect any usage, reset information or extra-usage setting the actual account displays. If a field or page is absent, record Not displayed and consult the provider’s current plan/help information; do not hunt guessed menus or upgrade for the exercise. Keep API billing separate from a chat subscription. Never write API keys or payment details on this card.