Standard
Seat + credits$30 per seat / month
500 credits
- Good fit for routine campus work with occasional advanced needs.
- Uses credits when you choose advanced capabilities.
License tiers
Effective July 1, 2026, OpenAI introduced a new billing model for Higher Education workspaces. In response, UMD has updated its ChatGPT licensing model to better align with these changes.
UMD ChatGPT licenses now use a credit-based billing model for advanced capabilities instead of a fixed pricing structure. Standard ChatGPT features remain included, while advanced capabilities draw from the monthly credit allocation associated with each licensing tier (see the attached overview of included and advanced capabilities).
Users may change their licensing tier at any time by submitting the ServiceNow ChatGPT License Request form. Please note that both the user and the designated financial reviewer must approve any upgrade request before it can be processed.
Power
More credits$100 per seat / month
2250 credits
- For teams that regularly use advanced reasoning or research tasks.
- More headroom for higher-cost modes and features.
Max
Highest pool$200 per seat / month
4750 credits
- Designed for heavy advanced usage and frequent Codex work.
- Best for sustained, high-volume or high-complexity workflows.
We understand that not everyone's usage patterns will fit neatly within these tiers. For that reason, if your initiative requires credits beyond those offered above, please contact us at dit-ais@umd.edu. We can work with you to establish an allocation that aligns with your anticipated usage and funding requirements.
Important- Credits DO NOT automatically rollover from month to month.
- Not every interaction necessarily consumes credits — core capabilities are included with your seat, and usage beyond included limits or use of credit-based features may consume credits according to the rate card.
Features & credit rates
Core capabilities are always included with your seat. Advanced features draw from your set amount of credit pool — the matrix below shows what's included, what's limited, and what costs credits every time.
Codex (token-based credit usage)
Codex credit usage is based on token type: input tokens, cached input tokens, and output tokens. The table below lists credit rates by model.
| Model | Input Tokens | Cached Input Tokens | Output Tokens |
|---|---|---|---|
| GPT-5.5 | 125 credits | 12.5 credits | 750 credits |
| GPT-5.4 | 62.50 credits | 6.250 credits | 375 credits |
| GPT-5.4-Mini | 18.75 credits | 1.875 credits | 113 credits |
Example calculation 1
Codex
If a GPT 5.5 Codex run uses 20,000 input tokens, 80,000 cached input tokens, and 5,000 output tokens, the run would consume about 7.25 credits:
- Input:
(20,000 ÷ 1,000,000) × 125 = 2.5 credits - Cached input:
(80,000 ÷ 1,000,000) × 12.5 = 1.0 credit - Output:
(5,000 ÷ 1,000,000) × 750 = 3.75 credits - Total:
2.5 + 1.0 + 3.75 = 7.25 credits
Example calculation 2
ChatGPT
After your daily free limits are used, an afternoon session with 2 extra image generations, 1 extra Deep Research task, and 4 GPT 5.5 Thinking messages would consume 100 credits:
- Image Gen — 2 over the 3/day free limit:
2 × 5 = 10 credits - Deep Research — 1 over the 5/day free limit:
1 × 50 = 50 credits - GPT 5.5 Thinking — 4 messages:
4 × 10 = 40 credits - Total:
10 + 50 + 40 = 100 credits
Codex speed note: Codex usage on Fast mode consumes credits about 1.5x faster than Standard mode.
Notes & links
Rollout notes and official references.
Official rate cards
OpenAI Help Center
