UAB restructures ChatGPT Edu into three credit-metered tiers, putting Work, plugins, Codex and MCP connections behind higher prices

University of Alabama at Birmingham (UAB) IT · Higher education · Birmingham, Alabama, United States

Publisher: UAB IT News
Original publication: October 30, 2026
Source retrieved: 2026-11-01

What happened

UAB will offer faculty and staff three ChatGPT Edu tiers: Basic (chat, 450 credits a month, $29.25), Intermediate (adds ChatGPT Work and plugins, 900 credits, $49.50) and Advanced (adds Codex and approved MCP server connections, 2,200 credits, $108). Existing subscribers move automatically to Intermediate, Codex subscribers to Advanced; Advanced users can buy up to three 1,000-credit packs at $45 a month. Access requires acknowledging UAB Responsible AI Use, and plugins and connections must pass UAB data-handling approval.

Why it matters

A public research university is metering agentic features (Work, Codex, MCP) by credits and tier, a concrete model for how campuses and agencies can budget and gate OpenAI agent capability rather than buying flat seats.

Evidence and measured results

Operator notice with specific prices and credit allotments. No usage, cost-per-task or adoption data; unclear whether the prices are internal chargebacks or OpenAI list pricing.

Limitations and uncertainty

Interpretation: credit allotments are not translated into task volumes, so budget adequacy is unknown. Students are excluded. Pricing applies to UAB only.

Put this evidence to work

Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-11-01; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.

Sales

Role takeaway

Customer problem: Campuses need a way to budget and govern agentic ChatGPT features without giving every user the most expensive, most connected tier.
Stakeholders: CIO and central IT
Department budget holders
Research computing
Information security office
Discovery questions: How many users actually need Codex or MCP connections?
How will you forecast credit consumption by tier?
Who approves an MCP server connection?
Value hypothesis: Tiering may match spend to need and keep agent connections behind review, if usage data supports the tier choices.
Potential engagement: A bounded credit-consumption and tier-fit analysis for one department over one billing cycle.
Unsupported claims: That tiering lowers total cost
That 2,200 credits covers typical Codex workloads
That MCP connections are secure by default

Pre-sales engineering

Role takeaway

Fit: Higher-ed and agency ChatGPT Edu/Enterprise workspaces offering Work, Codex and MCP.
Architecture and integration: Approved MCP servers connecting ChatGPT to institutional systems; plugins subject to data-handling review.
Prerequisites: MCP connection approval process
Responsible-use policy
Usage reporting
Constraints: Faculty and staff only
Approval criteria unpublished
Credit-to-usage mapping unknown
Security: Each MCP server is a new data path; require scoped credentials, logging and data-classification review per connection.
Proof of value: Measure credit burn for a sample of Codex tasks
Pilot one approved MCP connection with access logging

Delivery

Role takeaway

Work: Migrate subscribers to tiers, monitor credits, and run MCP/plugin approvals.
Dependencies: Billing data from OpenAI
Security review capacity
Ownership: Central IT with department budget holders.
Skills and adoption: Guidance on which tasks justify Work or Codex use and how credits are consumed.
Governance checkpoints: Security review per MCP connection
Quarterly tier-fit review
Acceptance criteria: Credit usage reported by tier monthly (proposed)
All MCP connections documented with approval record (proposed)
Risks: Credit overruns
Unreviewed data exposure via connections
User confusion over tier changes

Implementation considerations

Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.

Architecture and integration

What must connect, and where does the AI sit in the workflow?

MCP server connections are limited to the top tier and to UAB-approved servers, so agent access to institutional systems is an explicitly governed integration rather than a default.

Governance

Who approves, reviews and stays accountable for outcomes?

Ties capability tier to a responsible-use acknowledgment and to approval of plugins and data connections; a template for tiered agent governance.

Security and privacy

What data, permissions and controls need testing?

Codex and MCP connections widen data and system access; UAB relies on an approval process for connections, details of which are not published.

Accessibility and workforce

Who is affected, and what skills or accommodations follow?

Higher tiers may concentrate agent capability among departments able to pay.

Procurement

What should contracts, pricing and exit terms secure?

Shows a consumption-metered, tiered internal offering on top of an OpenAI Edu agreement; procurement should obtain credit-to-usage definitions.

Operating model

Which teams own the service once it runs?

Turns AI provisioning into an ongoing chargeback and approval service run by central IT.

Publication history

2026-10-31OpenAI · Issue 033 resources

Read preserved resource versions (JSON)

Stable resource ID: uab-chatgpt-edu-credit-tiers-codex-mcp-20261030