Model usage, limits, and Extra Usage

Your license includes every TextCortex feature, and there is no feature gating. However, premium AI models and extended agentic runs draw from a compute budget. This article explains how that budget works, what happens when you reach it, and how Extra Usage fits in.

Default token wallet

Every seat includes a default token wallet worth 50% of your monthly license fee in compute value. This wallet covers:

• Premium model queries
• Deep research and multi-step agentic tasks
• Daily AI workflows that run longer than standard configurations

If you use up your default wallet, you keep working without interruption. You revert to standard models and default platform settings. There are no lockouts and no surprise overuse bills.

Premium models vs. standard models

• Premium models deliver advanced reasoning, larger context windows, and frontier capabilities. Using them draws from your default wallet.
• Standard models handle everyday tasks and remain fully available even after your default wallet is empty.

Important: budget limits apply only to premium model usage. Standard models continue to work exactly as before.

Extra Usage (optional)

If you need more headroom for heavy workloads, your admin can enable Extra Usage. It sits on top of your license and is completely optional.

• Billing: Extra Usage is billed at the base large language model cost plus a platform markup. It only applies when Extra Usage is active.
• Spend order: You draw from your default wallet first. Once that is exhausted, you draw from your individual Extra Usage allowance.
• Admin controls: Admins set a central spend cap with full visibility, and can raise individual caps for power users.
• Soft landing: If you hit your Extra Usage limit, you simply revert to standard models and default configurations. Again, no lockouts.

What Extra Usage unlocks:

• Longer agentic runtimes (up to roughly 2 hours on a single task, compared with roughly 10 minutes under standard configurations)
• Adjustable context windows sized to the task
• Frontier models such as Claude Opus 4.7, which run only under Extra Usage

Optimize your token spend

The model you choose is the biggest lever on your budget. Kimi and GLM models deliver near-state-of-the-art performance at a much lower token burn than frontier-class models, and both support long context windows. If data residency matters, Kimi also offers EU-hosted inference (your admin can enable this in workspace settings). Choosing efficient models helps your default wallet stretch further.

Text Model Pricing (EUR per 1M tokens)

Image Generation Models

If you have questions about your current spend or need an individual cap adjustment, contact your workspace admin.