Skip to main content

ChatGPT Rate Card

Flexible pricing rate card for the Enterprise and Business plans

Updated over a week ago

Note: As of August 29th, 2025, we've renamed ChatGPT Team to ChatGPT Business. For more information and questions related to the name change, please refer to our article: ChatGPT Business Rename FAQ

This article outlines the current rates for ChatGPT models, features, and Codex under the flexible pricing structure. For more information about how flexible pricing works, please refer to our article: Flexible pricing for the Enterprise, Edu and Business plans

ChatGPT

Models

Unit

Credits (~)

1 message

N/A - Unlimited

GPT-5 Thinking

1 message

10 credits

GPT-5 Pro

1 message

50 credits

* To provide the best possible response, GPT-5 may occasionally auto-select GPT-5 Thinking for more complex tasks. When this happens, you will be charged at the GPT-5 Thinking rate. This will show up in the analytics dashboard as "gpt-5-auto-thinking".

Features and products

Unit

Credits (~)

1 message

30 credits

1 task

50 credits

1 message

5 credits

1 minute

5 credits

Legacy Models

Unit

Credits (~)

GPT-4o

1 message

N/A - Unlimited

GPT-4.1

1 message

2 credits

GPT-4.5

1 message

20 credits

o3

1 message

10 credits

o3-pro

1 message

50 credits

o4-mini

1 message

5 credits

Note: One prompt and response form a single message.

Codex

Unit

Credits (~)

Local Tasks

1 message

~5 credits on average (variable)*

Cloud Tasks

1 message

~25 credits on average (variable)*

Code Review

1 pull request

25 credits (fixed)

Note: Credits for Cloud tasks and Code Review will be waived until October 20th.

*Credits per message for local tasks and cloud tasks will vary based on the size, complexity, and reasoning required.

  • Local messages typically use around 5 credits per message, with most staying below 10 credits.

  • Cloud tasks typically use around 25 credits per message, with most staying below 60 credits.

Code Review is charged as a fixed fee per pull request when Codex runs reviews through GitHub — whether triggered by @codex review or enabled automatically across your repository.

Credit rates are the same when using either GPT-5 or GPT-5-codex.

Did this answer your question?