Ask questions, brainstorm, draft text, and summarize information.
Built for the way you work
UMD's ChatGPT EDU workspace provides access to the latest models enabled by UMD for everyday assistance, research, content creation, and coding.
Ask questions, develop ideas, and work through everyday tasks.
Find current information and review linked sources.
Use custom assistants configured for specific tasks. Custom GPTs are moving to plugins.
Draft and revise writing or code in an interactive workspace.
Explore concepts with guided, step-by-step learning support.
Investigate questions, analyze files, and create visual material.
Analyze documents and datasets, compare information, and explore patterns.
Create and edit visuals for documents and presentations.
Investigate questions across sources and produce reports for review.
Work through complex problems and carry multi-step tasks through to results you can review.
Work through complex questions that need deeper analysis.
Use more intensive reasoning for demanding problems.
Talk through ideas and tasks instead of typing.
Use tools to carry out multi-step tasks, with your review of important actions.
Create reports, presentations, analyses, and project plans through multi-step workflows.
Build and debug code, run tests, and review changes with developer-focused tools.
Available in UMD's workspace to approved users.
GPT-Rosalind, OpenAI’s reasoning model for life sciences research, is available in UMD’s ChatGPT EDU workspace. Access is limited to approved users, so it isn’t turned on for every account. To request it, email DIT AI Services.
Plans, credits, and limits
UMD offers four ChatGPT EDU subscription tiers with the same enabled features and models. Choose a tier based on the monthly credit allowance you need. Credits do not roll over, and rates can change.
- Good fit for routine work with occasional advanced needs
- Same enabled features and models as higher tiers
- For teams that regularly use advanced reasoning or research
- More room for higher-cost modes and features
- Designed for heavy advanced usage and frequent Codex work
- More room for sustained, high-volume work
- Best suited for sustained, research-intensive work
- For sustained advanced usage and frequent Codex work
Credit-free capabilities in Chat
Included allowances in Chat
These allowances apply only in Chat, not Work or Codex. Each user receives the following without consuming credits:
3 image generations
Chat onlyPer user, per rolling 24-hour period. Additional image generations in Chat consume credits after you use this allowance.
5 deep-research queries
Chat onlyPer user, per rolling 24-hour period. Additional deep-research tasks in Chat consume credits after you use this allowance.
Each feature's rolling 24-hour window begins when you first use that feature. The allowances do not reset at midnight.
How credits are used
Credit-consuming Chat usage
ChatReasoning, Pro reasoning, agent mode, and connected Voice usage consume credits. Image generation and deep research also consume credits once you use their included Chat allowances.
Work and Codex usage
Token-basedIn UMD's workspace, model processing in Work and Codex consumes credits from the first request. Usage depends on the model, context, output, reasoning effort, speed, and tools you use.
Work and Codex image generation and research do not draw from the included Chat allowances.
When you reach your limit
What you can still use
You can continue using zero-credit Chat capabilities and any remaining included image generations or deep-research queries, subject to their separate feature limits.
What pauses
Further credit-consuming requests are blocked. An Instant request that needs credit-consuming reasoning may also be blocked. Reaching your limit does not trigger a general automatic fallback to another credit-consuming model.
A request you start below your cap may finish and take you slightly above it before ChatGPT blocks further credit-consuming requests.
Check your usage dashboard to understand your remaining credits. To request a different monthly allocation, follow the existing subscription tier-change process.
Moving custom GPTs to plugins
OpenAI is replacing custom GPTs with plugins and will retire custom GPTs on December 11, 2026. Plugins work much like custom GPTs: they can package instructions and reference material as skills, connect to tools and external systems, or both. If you built GPTs that you or your colleagues rely on, migrate them to plugins before December 11.
Open My GPTs in ChatGPT and migrate the GPT to a plugin. Its instructions become a skill, its knowledge files are copied into the plugin’s reference files, and connected apps can be carried over.
What doesn’t carry over
- Existing conversations with the GPT.
- The model the GPT was set to use.
- Sharing settings: migrated plugins start private, so share the plugin again with the people who used your GPT.
- Custom actions, which may need to be rebuilt with an available app or a custom MCP server.
A plugin may respond differently from the original GPT, so test your usual prompts, reference material, and output formats before you share it.
OpenAI API projects
DIT tracks UMD’s OpenAI API projects in the DIT Hosting portal. To request a new project or a change to an existing one, submit the OpenAI API Project Request form.
Use the request form to update an existing project. Describe the changes or updates you need in the form’s note, such as:
- Driver worktag: the driver worktag your project’s API usage is billed to.
- Workspace membership: people to add to UMD’s OpenAI API workspace. Once they are in the workspace, project owners can add them to specific projects.
- Other project configuration.
- One driver worktag per project: every API key in a project is billed to that project’s driver worktag. Each project can use a different driver worktag.
- Workspace first, then projects: people must first be added to UMD’s OpenAI API workspace. Only workspace members can then be added to specific projects by project owners.
- Project owners manage their project’s team, adding and removing workspace members.
- Project members issue and revoke API keys and can view the project’s monthly usage and charges. Owners and members are responsible for keeping keys secure.
- Billing: API usage is billed by the tokens you use to the project’s driver worktag, separately from ChatGPT licenses. You don’t need a ChatGPT license to use the API.
Access information
ChatGPT EDU at UMD is available to faculty, staff, and researchers with an active UMD account.
Use the ChatGPT License Request to request a new license or change your subscription tier. A tier change requires your approval and a designated financial reviewer's approval.
Sign in to ChatGPT using your approved UMD account. Explore the tools above to find a starting point for your work, and review UMD's data-classification guidance before uploading files or entering university information.
Use ChatGPT with the right UMD safeguards
ChatGPT EDU is approved for UMD Level 3 (High) data. OpenAI does not use ChatGPT EDU workspace content to train its models by default, and UMD retains rights to content submitted to the workspace. Follow UMD privacy, security, and records requirements and the settings enabled in your workspace. Review UMD data-classification guidance before entering university information. Level 4 data requires additional review.
- Review generated content before using it
- Verify facts, citations, calculations, and recommendations
- Keep a qualified person responsible for decisions
- Follow course, research, departmental, and university requirements
- Enter restricted or highly regulated information without the required review
- Treat ChatGPT as a final authority or a substitute for required judgment
- Use Level 4 data without additional review
Questions about ChatGPT at UMD? Contact DIT AI Services for help with licenses, access, and usage.
↑ Back to top