Introduction
What is Lumen?
Lumen is an AI model gateway for research institutions. It sits in front of many different AI models — hosted on different servers and platforms across your institution — and gives you a single, unified way to reach all of them.
You can interact with those models through a browser-based chat interface, or point your own applications and tools at Lumen's API and access every model the same way, without needing a separate account or key for each one. A single Lumen API key is all you need.
Tokens and Coins
Two numbers matter when you use Lumen: tokens and coins.
Tokens
A token is the basic unit of text that an AI model processes. Roughly speaking, one token is about four English characters or three-quarters of a word. The sentence "Explain quantum computing in plain terms" is about 8 tokens.
Every AI request involves two kinds of tokens:
| Kind | What it counts |
|---|---|
| Input tokens | Everything you send — your message, any files, and the conversation history |
| Output tokens | The model's reply |
Tokens are how the AI industry measures usage. Lumen tracks them so you can see how much of each model you are consuming.
Coins
Coins are Lumen's internal currency. They exist so that one budget can cover many different models — each with different real-world prices — without you having to think in dollars and cents.
Each model has a published rate in coins per million tokens (shown separately for input and output). When you send a message, the coin cost is calculated like this:
cost = (input_tokens / 1,000,000 × input_rate)
+ (output_tokens / 1,000,000 × output_rate)
Your coin pool is your budget. As you use models, coins are deducted from that pool. If your institution auto-refills pools, the balance tops up on a regular schedule.
Example: Ask a model a simple question — the model might use ~1,000 input tokens and ~300 output tokens. If the rates are 0.5 coins per 1M input and 1.0 coins per 1M output, the cost is:
(1,000 / 1,000,000 × 0.5) + (300 / 1,000,000 × 1.0)
= 0.0005 + 0.0003
= 0.0008 coins
You'd need over 1,000 similar questions to spend 1 coin.
In practice: you rarely need to think about the math. The Usage page shows your balance and burn rate, and the chat interface will warn you before you run out.
What You Can Do with Lumen
| Feature | Where |
|---|---|
| Chat with AI models in a browser | Chat |
| Track your token and coin usage | Profile |
| Create API keys for programmatic access | Profile → API Keys |
| Use Lumen from your own code or tools | API Reference |
| Browse available models and their capabilities | Models |
| Manage application clients (managers/admins) | Clients |
| Configure Lumen (admins only) | Admin Guide |
About Illinois Computes
Lumen was developed by the National Center for Supercomputing Applications (NCSA) as part of Illinois Computes.
Illinois Computes is a commitment from the University of Illinois Urbana-Champaign and the University of Illinois System that funds NCSA to provide — at no cost — computing, storage, and other technology resources, alongside consulting expertise, to researchers across campus and the system.
Illinois Computes' mission is to democratize access to cutting-edge technology and expert support for University of Illinois researchers across all disciplines and domains and all levels of computing experience. It is led by NCSA, drawing on its 40 years of experience supporting research computing nationally.
Illinois Computes supports all disciplines and is committed to meeting researchers where they are, whether they are seasoned experts in research computing or completely new to the idea of using technology to advance their research.
Feedback & Support
Questions and feedback are welcome. Send an email to help@ncsa.illinois.edu with Lumen in the subject line.
To report a bug or request a feature, open an issue on the Lumen GitHub repository.