Codex pricing: every ChatGPT plan and the API rates

Codex price on every ChatGPT plan

There's no standalone Codex subscription: you get it through a ChatGPT plan, and ChatGPT Work and Codex share the same pricing, credits and usage limits. Plus is the cheapest plan with Codex outside the desktop app.

Plan limits in brief

On Plus and Standard Business, OpenAI estimates you get 15 to 160 local messages every five hours on GPT-6.1 Sol, or 350 to 3,000 on GPT-6 Luna, though these are estimates rather than fixed caps and cloud tasks eat into your allowance faster. Pro plans currently have no five-hour Codex limit, and OpenAI notes that weekly limits may also apply. See the full breakdown of Codex usage limits for every model and plan.

Credits, Fast mode and Ultrafast

• When you hit your plan limit on Plus or Pro, you can buy ChatGPT credits to keep working without upgrading, with credit purchase prices that depend on your plan.
• At Standard speed, GPT-6.1 Sol costs 50 input credits and 250 output credits per 1M tokens (2.5 for cached input), and GPT-6 Astra costs 250 input and 1,250 output (25 cached).
• Fast mode burns through your included usage at 2.5x the Standard rate and uses purchased credits at 2x.
• Among Pro tiers, Ultrafast mode is only on Pro 500 (eligible Enterprise and Edu plans have it too), and on GPT-6 Astra and GPT-6.1 Sol it burns included usage at 8x and credits at 6x.
• Image generation drains your included limits about 3 to 5x faster on average, and voice in the desktop app draws from your usage budget at $0.05 a minute (or 1.25 credits a minute).

Codex on the API

If you'd rather pay per token than use a subscription, you can run Codex with an OpenAI API key and pay API token prices.

Signing in with an API key runs Codex in the CLI, SDK and IDE extension using the models available on your key, without cloud features like Slack or GitHub code review. Prompts over 272K input tokens get billed at long-context rates: gpt-6.1-sol jumps to $4 input and $15 output per 1M tokens, and gpt-6-astra goes to $20 input and $75 output (GPT-6.1 Sol usage drawn from a subscription has no long-context surcharge in Codex). Check the OpenAI API pricing page for the full price list.

OpenAI warns that API token rates are separate from subscription usage and shouldn't be used to estimate how many tasks a plan includes. And if you're on ChatGPT Pro, the $100, $200 or $500 in monthly API credits you can claim won't cover Codex usage through API keys.

The table below lists Standard tier prices per 1M tokens for prompts up to 272K input tokens.

Which Codex plan to pick

• Trying Codex: Start on Free ($0) or Go ($8 a month) to test GPT-6 Luna at Standard speed in the desktop app, subject to rollout.
• A few focused sessions a week: Get Plus at $20 a month for GPT-6.1 Sol and GPT-6 Luna across the CLI, IDE extension, web and iOS, plus cloud integrations.
• Daily heavy use: Pick Pro ($100, $200 or $500 a month) to drop the five-hour limit and get expanded usage (Pro 100 and Pro 500 now include twice their previous Work and Codex allowance), and go with Pro 500 if you want Ultrafast mode; our ChatGPT Plus vs Pro guide compares the tiers.
• Teams: Business costs $25 a seat monthly or $20 a seat billed annually (for 2 or more users), doesn't train on business data by default, and adds larger virtual machines for cloud chats along with SAML SSO and MFA.
• CI and automation: Use an OpenAI API key in the CLI or SDK so you pay per token instead of managing seat limits.

Use your ChatGPT plan as an API

Plan2API connects a ChatGPT plan that includes Codex, like Plus or Pro, using a one-time device code and gives you a single p2a_ key for the OpenAI Chat Completions endpoint (POST /v1/chat/completions) at https://api.plan2api.com/v1. It reads your account's live Codex models and exposes them as openai/<model> (such as openai/gpt-6.1-sol, openai/gpt-6-astra and openai/gpt-6-luna), with streaming, tool calling, JSON mode and reasoning_effort supported. Read more about how the ChatGPT subscription API works.

Every request runs against your ChatGPT plan's own Codex limits, and Plan2API never bills per token or raises those limits. It speaks Chat Completions only, so there's no /v1/responses endpoint. Plan2API's Free plan ($0, no card needed) supports 1 connected subscription and 50 chat requests a day, while Pro is $12 a month (or $99 a year) with 3 subscriptions and no daily request cap.

curl https://api.plan2api.com/v1/chat/completions \
  -H "Authorization: Bearer $PLAN2API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-6.1-sol",
    "reasoning_effort": "medium",
    "messages": [
      {"role": "user", "content": "Write a bash one-liner that counts lines in all .py files"}
    ]
  }'