tokens-per-dollar · prices read

Codex usage: what your week costs on Plus, Pro and the API

ChatGPT Plus is $20 a month and Pro starts at $100 (tiers at $100, $200 and $500 a month, billed monthly); Codex credits and the OpenAI API bill per token, and at $0.04 a credit (derived from OpenAI's own figures) every rate on the rate card comes to the model's API list price. Which is cheapest for you depends on how much of a week your own work fills, and Codex records that: npx --allow-git=root github:agentwares/tokens-per-dollar forecast reads its readings and your last 14 days and prices the same pace on all four. On the plan you are on, the share of your week its limit holds is measured from Codex's readings and labelled an estimate once carried forward; on a plan you are not on, OpenAI publishes no limit to measure against, so the page and the command give the condition instead of a number.

$ npx --allow-git=root github:agentwares/tokens-per-dollar forecast demo
  A 30-day month at this pace
    ChatGPT Plus   $20 + $8.80  its weekly limit holds 74% of your week (estimate); the rest is
                                about 220 credits a month at $0.04 a credit (derived)
    ChatGPT Pro      from $100  no 5-hour limit (tiers at $100, $200 and $500 a month, billed
                                monthly); the fee alone carries your week if its weekly limit is at
                                least 1.4× ChatGPT Plus's, which OpenAI does not publish
    Codex credits       $33.56  every token at the credit rate card, $0.04 a credit (derived)
    OpenAI API          $34.76  every token at list prices
Synthetic history: the real output of npx --allow-git=root github:agentwares/tokens-per-dollar forecast demo on generated transcripts, with Codex weekly readings and Claude Code limit refusals added in their real formats; not anyone's usage (as of 7 October 2026).

The four ways to pay for Codex

PlanPer monthHow your week is priced
ChatGPT Plus$20measured from Codex's own readings when you are on it: how much of your week its weekly limit holds, and the rest at the credit rate card
ChatGPT Pro$100 (tiers at $100, $200 and $500 a month, billed monthly)its fee, and the multiple of your plan's weekly limit it would need to hold your week; OpenAI publishes neither as a number
Codex creditspay as you goevery token at OpenAI's credit rate card
OpenAI APIpay as you goevery token at the API list price

ChatGPT Plus: a five-hour window and a weekly limit; OpenAI publishes local-message estimates per five hours that it calls not fixed limits; extra credits can be bought past the limit. ChatGPT Pro: no five-hour limit; weekly limits, not published in tokens or as a multiple of Plus; extra credits can be bought past the limit. Past the limit on either, you can buy credits at the rate card, which is how the forecast prices the part of your week a plan does not hold.

A usage dashboard that looks forward

Codex's own /status shows how much of each window you have used now, and trackers chart where you have been. What neither tells you is where your pace takes you this week and what that pace would cost on the plan you are not on. The forecast reads the same readings /status shows, sizes the weekly limit in credits from your own work, and answers both, in your terminal, without an account or a network request.

For when, rather than how much: will I hit my Codex weekly limit? And the rate card itself, per model: Codex credits to tokens.

Changelog: what the vendors changed, and when

Dates are the vendors' own. “As reported” means the press quoting the vendor, because the vendor's own post could not be read. Each entry says what changed, never why.

Run it on your own logs

On the machine where you use Claude Code or Codex:

$ npx --allow-git=root github:agentwares/tokens-per-dollar forecast
$ npx --allow-git=root github:agentwares/tokens-per-dollar

forecast reads the limit readings Codex keeps and your last 14 days of work, and says when your pace reaches the 5-hour and weekly limits and what that pace costs on Plus, Pro, Codex credits and the API, each projection labelled an estimate. With no command it prices your last 30 days on every plan. Free and MIT; it makes no network request and sends nothing. --json prints every price with its source and the date it was read.

For a team: seatledger keeps everyone's coding-agent ledger in one place and flags when the token rate moves. Free for up to 3 developers; email alerts on Team, $49 a month.

Sources