Codex Pricing

Eligible users can send Codex invitations from the profile menu in the lower-left corner of the app. Choose Invite a friend on an eligible personal plan or Invite a coworker in an eligible Business workspace, enter the recipient's email address, and send the invitation.

The invitation dialog shows the current reward, recipient requirements, invite limits, and when rewards expire for the plan or promotion. Personal and Business referral programs use different rewards and eligibility rules. Referrals are not currently available for ChatGPT Enterprise.

The official source records a June 11–24, 2026 invitation promotion for eligible Plus and Pro users; that promotion has ended. During the promotion, eligible users could invite up to three friends. When an eligible recipient sent their first Codex message, both people received a banked rate-limit reset usable for 30 days after it was granted. Business referrals use separate shared-workspace credit rewards; review the current terms before sending an invitation.

How much does Sites cost?

Sites is included with eligible ChatGPT plans during public beta. Availability depends on the plan, region, and workspace settings.

What are the usage limits for my plan?

The number of messages you can send depends on the model, the size and complexity of the task, and whether it runs locally or in the cloud. A small script or routine function may use only a fraction of the allowance, while a larger project, long-running task, or extended session that requires the agent to hold more context can use substantially more per message.

Tasks that look similar can consume different amounts. Model choice, context, reasoning, tool use, retrieval, and caching all affect usage, so prompt length alone is not a reliable estimate.

Choose the GPT-5.6 model that fits the work:

• Sol is built for the hardest work—complex reasoning, ambiguous problems, advanced coding, and high-stakes decisions.
• Terra is the everyday workhorse for production tasks, reporting, document analysis, coding, and work that requires sound judgment.
• Luna is optimized for fast, high-volume work such as routing, classification, extraction, support, background automation, and focused coding tasks.

The following ranges are local-message estimates from the official source snapshot. Values are messages in a shared five-hour window; actual usage varies with the task.

Local messages and cloud chats share a five-hour window, and additional weekly limits may apply. The model rows in the source snapshot mark Cloud chats and Code Reviews as Not available; use the account usage dashboard and the relevant feature page as the source of truth for current product availability.

Enterprise and Edu users with flexible pricing do not have fixed rate limits; usage scales with credits. Enterprise and Edu plans without flexible pricing have the same per-seat limits as Plus for most features.

Usage limits are shared with other agentic features after pricing for those features takes effect. This currently includes ChatGPT for Excel on Plus and Pro.

Speed configurations increase credit consumption and therefore use included limits faster. Fast mode consumes credits at a higher rate on supported models; see Speed. Image generations also use included limits about 3–5 times faster on average than comparable turns without image generation, depending on image quality and size.

GPT-5.3-Codex-Spark is a research preview for ChatGPT Pro users only and was not available in the API at launch. It runs on specialized low-latency hardware and has a separate usage limit that may change with demand.

ChatGPT Voice in Desktop

ChatGPT Voice on desktop uses a separate, plan-dependent allowance measured in rolling five-hour windows. Tasks started through Voice use your existing Codex usage budget. ChatGPT notifies you when you reach either limit.

ChatGPT Voice in Desktop uses a duplex model: GPT-Live manages the live conversation, while GPT-5.6 Terra starts and coordinates tasks in the app.

• Plus: Approximately 15–30 minutes.
• Pro 5x ($100/month): Approximately 1–2.5 hours.
• Pro 20x ($200/month): Unlimited voice access.
• Business: Approximately 45 minutes.
• Enterprise / Edu (legacy): Approximately 45 minutes.

Unlimited voice access doesn't make Codex tasks unlimited. Tasks started through ChatGPT Voice continue to use your existing Codex usage budget.

For Business, Edu, and Enterprise workspaces with credit-based or pay-as-you-go billing, Desktop voice costs approximately 6 credits per minute. ChatGPT Voice in Desktop is not available via API key currently.

What happens when you hit usage limits?

If you reach a usage limit during an active turn, the agent can continue working on that turn, subject to fair-use limits.

ChatGPT Plus and Pro users can purchase additional credits after reaching the included limit without changing plans. Business, Edu, and Enterprise plans using flexible pricing can purchase additional workspace credits.

When you are close to a limit, you can also switch to a smaller model. All users can run additional local chats with an API key and pay standard API rates.

How does image generation count toward usage limits?

Image generation uses the same general usage limits as local messages and cloud chats. Depending on image quality and size, it uses included limits about 3–5 times faster on average than similar turns without image generation. After the included limit is exhausted, image generation also draws from credits.

Image generation is not available on the Free plan. When Codex uses an API key, API pricing applies to image generation instead of the ChatGPT plan's included limits.

Where can I see my current usage limits?

Use the usage dashboard to see current limits. In an active Codex CLI session, use /status to view the remaining allowance.

Check the dashboard every week or two to understand consumption and remaining capacity. If use is higher than expected, consider whether a smaller model or a more focused task scope still produces a useful result.

What are tokens and credits?

Tokens are small units ChatGPT uses to read and write information. Prompts, files, chat history, tool results, and responses all consume tokens.

Credits convert token use into a unit that is easier to track and manage. Credit cost varies with the model, context, reasoning, and tools. After included plan usage is exhausted, work can continue while credits remain.

Usage is calculated in credits per million input, cached-input, and output tokens. Learn more about tokens.

The official source lists the following credit rates per million tokens for models and features:

GPT-5.6 uses an average of 5–40 credits per message. Fast mode uses a higher credit rate for supported models; see Speed for current rates.

A small number of Enterprise customers should continue using the legacy rate table until they migrate to the new token-based pricing. Contact OpenAI Sales if you need confirmation.

• Learn about credits in ChatGPT Plus and Pro
• Learn about credits in ChatGPT Business, Enterprise, and Edu

What counts as Code Review usage?

Only reviews that Codex runs through GitHub count as Code Review usage, such as requesting a review with @Codex on a PR or enabling automatic reviews for a repository. Reviews run locally or outside GitHub count against general usage limits.

What can I do to make my usage limits last longer?

The usage and credit figures above are averages. To use the allowance more efficiently:

• Control prompt size. Keep instructions for the agent precise and remove irrelevant context.
• Limit source material. Provide only relevant files and narrow sources or date ranges when possible.
• Match the output to the actual need. Specify the audience, format, and length, and separate required work from optional polish.
• Reduce AGENTS.md size. In large repositories, layer AGENTS.md files so only relevant project instructions enter the context.
• Limit MCP servers. Each MCP server adds context and usage; disable servers that are not needed.
• Choose a smaller model for routine work. Depending on the current model, GPT-5.6 Terra or GPT-5.6 Luna may extend the local-message allowance.

For more guidance on choosing and scoping tasks, see Use Work efficiently.

Features can depend on the plan, region, account, workspace policy, and rollout stage. Items marked * are currently available only in selected regions; † means some first-party plugins are unavailable. Use the feature documentation and the state shown in the account as the source of truth.

Access and surfaces

Models and multimodal

Local features

Browser and remote control

Customization and extensions

Cloud and integrations

Admin, security, and analytics