🪙Tokens and Limits

What are tokens and what are their limits?

Tokens are essentially "points" that are used for each request made in the bot. They are spent on tasks such as text processing, image generation, and selecting different models (like GPT-4o Mini, GPT-4o, o1, and Claude Sonnet 3.5).

Each day, users receive 100,000 free tokens. If you reach the daily limit, you can purchase additional tokens in the "Balance" (Subscription) section, in any amount you need. This is especially useful if you need a large number of tokens at once, such as when writing a book.

There is also an option to earn bonus tokens for free: by inviting a friend with a referral link, you receive 10,000 tokens. Go to the "Get Free Tokens" section in the menu to access this option.

Token Consumption Calculation System

In our system, a differentiated approach is used to calculate the cost of requests depending on the selected model. Each model has its own unique multiplier that determines how many tokens will be consumed when using a specific model.

How is the cost calculated?

The cost of a request is calculated using the following formula:

Total Cost = Number of Tokens in Request × Model Multiplier

For example:

• If your request contains 1000 tokens and you use the gpt-4o-mini model (multiplier 1), the total cost will be 1000 tokens .

• If the same request (1000 tokens) is processed using the gemini-pro-1.5 model (multiplier 8), the total cost will be 8000 tokens .

Model Multipliers Table

Model	Multiplier
GPT 4o Mini	1
Deep Seek	1
Gemini Flash 1.5	1
Gemini Pro 1.5	8
GPT 4o	15
Claude 3.5 Sonnet	20
Claude 3.5 Haiku	8
GPT o1 Preview	100
GPT o1 Mini	20

Examples of Calculations

1. Simple Request :

   • Request: 500 tokens

   • Model: gpt-4o-mini (multiplier 1)

   • Total Cost: 500 × 1 = 500 tokens

2. Complex Request :

   • Request: 750 tokens

   • Model: google/gemini-pro-1.5 (multiplier 8)

   • Total Cost: 750 × 8 = 6000 tokens

3. High-Performance Model :

   • Request: 200 tokens

   • Model: o1-preview (multiplier 100)

   • Total Cost: 200 × 100 = 20,000 tokens

Important Notes

• The higher the performance and complexity of the model, the higher the multiplier.

• When choosing a model, consider not only its capabilities but also the cost of usage.

• For testing purposes, it is recommended to use models with low multipliers, such as gpt-4o-mini or gemini-flash-1.5.