Skip to content

plan-allotments.md

Latest commit

 

History

History
52 lines (37 loc) · 3.38 KB

plan-allotments.md

File metadata and controls

52 lines (37 loc) · 3.38 KB
Raw

ChatGPT plan allotments

Reference map from plan_type to plan name, seat weight, and monthly price, as implemented in lib/plan-allotment.ts and covered by test/plan-allotment.test.ts. Update all three together.

plan_type arrives from two places that agree: the chatgpt_plan_type claim in the OAuth access token (named via lib/auth/plan-tier.ts) and the plan_type field on /wham/usage (read live by codex-limits and the TUI).

Normalization

normalizePlanSlug reduces a raw slug before matching: lowercased, a leading chatgpt prefix stripped, _/-/spaces folded to single spaces, and a trailing plan word dropped. So team, chatgptteamplan, and ChatGPT_Team_Plan all land on team.

The map

Accepted plan_type (normalized) Plan Weight Multiplier Monthly (USD)
plus ChatGPT Plus 1 1x 20
team ChatGPT Business (old slug, still emitted) 1 1x 25
business … standard (e.g. business_standard) ChatGPT Business Standard 1 1x 25
business … prolite / pro lite / premium (e.g. self_serve_business_prolite) ChatGPT Business Premium 5 5x 125
pro, pro 20x, pro 200 ChatGPT Pro 20 20x 200
prolite, pro lite ChatGPT Pro Lite 5 5x 100
pro 5x, pro 100, pro legacy, legacy pro, legacy pro 5x, pro legacy 5x ChatGPT Pro (legacy $100 tier) 5 5x 100
business (bare) Business workspace, seat unstated — — —
free ChatGPT Free — — —
go ChatGPT Go — — —
enterprise ChatGPT Enterprise — — negotiated per contract
anything else unrecognized — — —

Explicit-match notes:

  • team is matched before the business matcher because OpenAI kept the slug while renaming the product to Business.
  • self_serve_business_prolite is the premium Business seat ($125), not the personal Pro Lite tier ($100) that shares the prolite token — the business word in the slug decides.
  • The six legacy spellings are the $100 Pro that predates the $200 Pro; they are matched before pro claims the bare word.
  • A bare business names the workspace, not a seat, and its two seats are 5x apart, so no ratio is stated.

The math

  • The weight is OpenAI's published per-seat ratio against the 1x seat, taken from monthly price (Pro $200 vs Plus $20, marketed as 20x). It describes the subscription, not a measured token allowance.
  • Every pool figure is a weighted mean over the summed weights, not a plain average: a Pro seat spent to 50% gives up twenty times what a Business Standard seat does at 50%.
  • A plan with no ratio (— rows) counts as one baseline seat in the total (DEFAULT_PLAN_WEIGHT = 1) so it cannot remove the other accounts, and prints no Nx badge rather than asserting a 1x OpenAI never published.

Where it shows up

  • Pool status line (quotaStatus.mode: "overview"; quotaStatus.multipliers prints the badge): 24%: #1 5x 13%, #2 20x 100% 3d, #3 1x 12%
  • codex-limits / limits CLI: Plan: Pro (20x) per account, closing with Pool: 93% used of 81x across 11 accounts — the 81x is the summed weight.

Both read the same module, so the two figures cannot drift apart.

Related