1. Check Your Usage Regularly
Check your credit usage periodically—especially before starting a large or complex task. This can help you identify unexpectedly high consumption while you still have time to adjust your model, reasoning effort or approach.
To view your current usage and monthly credit budget:
- Sign in to ChatGPT using your Clemson account.
- Confirm that you are in the Clemson ChatGPT Edu workspace.
- Open your profile menu and select Settings.
- Select Usage (or Usage & Billing in the Desktop App).
Note that ChatGPT may also display a 5 hour or Weekly limit; these are not currently enforced in our Clemson workspace and should be ignored.
2. Choose Your Model Carefully
Model choice can dramatically affect the number of credits consumed. Use the lowest-credit model that meets your task’s requirements. Reserve higher-credit models for work that requires their additional capability.
- Standard (default) model: Use for most work, including all ordinary ChatGPT conversations. These chats do not consume credits.
- Mid-level models: Use for everyday work requiring additional judgment, planning or tool use.
- Most-capable advanced models: Reserve for difficult, ambiguous or especially complex work requiring deeper analysis or greater polish.
Model names, availability and credit rates change over time. Consult the current OpenAI rate card to compare available models before beginning credit-intensive work.
3. Choose Your Reasoning Effort Carefully
Reasoning effort can also substantially affect credit consumption. Higher reasoning settings allow a model to spend more processing on a task. Use the lowest reasoning effort that produces a reliable result.
-
- Low or Light: Appropriate for quick, well-defined tasks.
- Medium or Default: Appropriate for everyday tasks requiring some planning.
- High or Extra High: Appropriate for difficult, multi-step tasks involving several sources or trade-offs.
- Max or Ultra: Reserve for exceptional cases in which depth is more important than speed or credit use.
Start with a lower or default reasoning setting. Review the result and increase the setting only when the task requires deeper planning, analysis or verification.