Go Tokens: Cost and Time Calculations

Estimate Go costs and run time with simple formulas: $1,000 = 10M Go Tokens and 1M tokens roughly equals 1 hour of non-parallel work.

Use Go Tokens to estimate what a workflow costs and how long it takes.

What are Go Tokens?

Go Tokens measure how much text an AI model processes while it runs your agent. Token usage drives the AI portion of your Go costs.

How much do Go Tokens cost?

Use this conversion when you estimate costs:

  • 10,000,000 Go Tokens costs $1,000
  • 1,000,000 tokens cost $100

How do you estimate time from token usage?

Use this rule of thumb for non-parallel work:

  • 1,000,000 Go Tokens take about 1 hour to produce on average
  • That equals ~60 minutes of non-parallel agent work per 1,000,000 tokens

This estimate helps you translate token usage into wall-clock time when a workflow runs steps one after another.

What does “non-parallel work” mean?

  • Non-parallel: one run or one step waits for the previous step to finish.
  • Parallel: multiple runs (or branches) execute at the same time.

Parallel execution often reduces wall-clock time, but it does not reduce token consumption. It can increase how fast you consume tokens because more work happens at once.

How do features change token usage?

Token usage depends on the text you send to the model and the output you request. A few Go features can raise token usage:

  • Knowledge Hub files: If a file from the Knowledge Hub acts as an input, it consumes tokens like other inputs.
  • AI Citations (grounded outputs): A grounded property costs about 20% more tokens than a non-grounded property.

Worked examples

Example 1: Convert tokens to dollars

You expect a workflow to use 2.5M tokens.

  • Cost estimate: 2.5M tokens * ($100 / 1M tokens) = ~$250

Example 2: Convert tokens to time (non-parallel)

You expect a workflow run to use 750k tokens.

  • Time estimate: 0.75M tokens * (60 minutes / 1M tokens) = ~45 minutes

Example 3: Convert budget to tokens and time

You have $1,000 for AI processing.

  • Tokens: 10M tokens
  • Non-parallel time: 10M tokens * (1 hour / 1M tokens) = ~10 hours

FAQs

Where do you check token usage?

Open Settings and review your Billing and Usage

Do tokens reset?

Reset behavior can vary depending on your plan or contract. To check your specific reset date, go to Billing and Usage for your plan details in Go, under Go tokens you’ll see a message indicating when your tokens will reset (for example, “Tokens will reset on…”).

Why does my time estimate not match actual run time?

This estimate averages across models, prompts, file sizes, and workflow design. Large inputs, longer outputs, citations, and Knowledge Hub context often increase tokens and time. Parallel execution changes wall-clock time but keeps token usage tied to total work.