Skip to main content

Compute Points

What are Compute Points?

Compute points are a metric used to quantify the consumption of computing resources when tasks are performed by our AI agents. Each task that you execute, whether it is sending a chat message, running an automation, or any other interaction with AI models, will consume a certain number of compute points.

The number of compute points consumed is directly related to:

  • The complexity of the task being performed.
  • The size and type of the language model (LLM) utilized.

Compute points allow us to manage and allocate our computational resources efficiently, ensuring all users have access to high-quality services while also tracking usage for billing purposes.

How Compute Points are Measured

The measurement of compute points is tied to the "tokens" used by language learning models (LLMs). In the context of LLMs, a token typically represents a piece of text such as a word or part of a word.

Here's how compute points correlate with token usage:

  • Lighter models, such as GPT-3.5 and Mistral Small, consume approximately 1-5 points per request.
  • More robust models like GPT-4 and Claude-3 may use between 10-30 points per interaction.

Please note that these figures are estimates and actual consumption may vary depending on the specifics of each request.

Compute Point Usage Examples

To help you better understand how compute points work, here are some examples:

Example 1: Simple Chat Message

Sending a straightforward chat message that requires only a minimal response from a lighter model like GPT-3.5 might consume around 2 compute points.

Example 2: Complex Automation Task

Running a complex automation task that involves generating lengthy content or processing extensive data through a more robust model like GPT-4 could consume upwards of 25 compute points.

Monitoring Your Compute Point Usage

Users can keep track of their consumption through our user dashboard. Here's how you can monitor your usage:

  1. Log in to your user dashboard.
  2. Navigate to the "Usage" section.
  3. Here, you'll find detailed reports on your compute point consumption.

It's important to regularly monitor your usage to ensure that it aligns with your expected compute point budget and to avoid any surprises on your bill.

Conclusion

Compute points provide a transparent way for users to understand and manage their use of computing resources when interacting with AI agents. By monitoring your usage and being aware of how different tasks consume varying amounts of compute points, you can optimize your operations accordingly.

For more information or if you have any questions regarding compute points, please refer to our FAQ section or contact our support team.