Skip to main content
All CollectionsAI in Tines
AI action credits & executions
AI action credits & executions

This article walks through the models available today for the AI action and the execution credits associated with them.

Kelli Hinteregger avatar
Written by Kelli Hinteregger
Updated over a month ago

The AI Action in Tines is limited by credits available. All plans, including our community edition, receive an allocation of monthly credits free.

Included credits

Edition

Monthly credits

Community Edition

50

Paid Editions

5000

Credits do not roll over from the previous month. Paid editions of Tines can purchase AI credit add-ons (100K monthly credits) through your Customer Success Manager.

Cost per execution

Credits are deducted based on the model selected for the AI Action. The amount of credits used for an execution reflects the relative costs incurred by Tines for the execution.

LLM

Credits per Execution

Haiku | Anthropic Claude 3

1 credit

Sonnet | Anthropic Claude 3

12 credits

Llama 3 70B | Meta

10 credits

What is an execution?

An execution is when the AI Action emits an event. We do not charge you based on the size of the prompt.

Max token sizes

While we don't charge for inputs there are natural limits implemented for inputs:

  • Input token limits

    • Claude 3: 20K

    • Llama 3: 8K

  • 4096 output tokens

For optimal performance, we do rate limit credits by action runs per tenant per minute. See our docs for details on rate limits.

What if I run out of credits?

You trust us with your most important workflows, we will protect those workflows and provide you with warnings when you’re nearing your credit limit. You can stay on top of your credit usage in Reporting.

How do you keep tabs on your credit limit?

In the AI settings modal, you can see your credits used to date by each team.

Understanding the available models

The AI Action supports 3 LLMs and counting. We’re expanding our model support as AWS Bedrock expands its support. Today, we offer the Anthropic Claude 3 and Meta Llama 3 70B models. Currently, Llama 3 is only available on US tenants.

Rate limits

For optimal performance, we do rate limit credits by action runs per tenant per minute. These limits vary by region:

  • EU tenants: 40 runs per tenant per minute

  • US-tenants: 100 runs per tenant per minute

Tip! To check where your tenant is hosted, append /info to your tenant URL.
The URL should look like this: https://acme-corp.tines.com/info

Did this answer your question?