Alex Models and Limits

The message limits and premium AI models system allows you to customize Alex to your needs.

Message Limits

How Counting Works

• Counted: Your queries and commands
• Not counted: Alex’s responses, sub-agent outputs
• Example: “Hello” = 1 message, “Try this” = 1 message |

Limit Indicator in Chat

The chat displays an indicator with remaining messages:

Note: When the limit is exhausted, a banner appears with Premium upgrade or Boost options.

Premium Upgrade

Extended limit for active users.

Important: Boost adds to the base limit. Premium does not renew automatically — after 30 days you return to the base limit.

AI Models

Model Comparison

When to Use Which Model

Sonnet: Quick queries, plugin installation, basic configuration, simple scripts.

Opus: Complex debugging, architectural decisions, complex scripts, multi-file analysis.

GPT-4 Turbo: If you prefer OpenAI models or need specific GPT-4 capabilities.

Tip: The basic model is sufficient for most tasks. Premium models are worth it for complex problems.

Token Billing

Premium models use a token-based system:

1. Pre-authorization: Before each message, the system estimates costs and blocks the amount from your credit
2. Actual consumption: After the response, actual token consumption is calculated
3. Difference refund: If the estimate was higher, the difference is returned

Example:

Estimated cost: 10,000 tokens × 0.003 CZK = 30 CZK (blocked)
Actual consumption: 8,500 tokens × 0.003 CZK = 25.50 CZK
Refunded: 4.50 CZK

Note: For premium models, you need sufficient credit on your account for pre-authorization.

Switching Models

1. In the chat, click the settings icon (gear) in the header
2. Select the desired model
3. The change applies immediately to the next messages

Next Steps

• Alex AI — Introduction to Alex
• SSH Setup — SSH access for VPS
• FAQ — Frequently asked questions

Need help? Open a support ticket  or ask Alex in your panel.