Equivalent API $ and Value ×
Default agent turn: 12,000 input tokens with 70% cache hit plus 1,500 output tokens. Unit prices prefer OpenRouter fair listings (snapshot 2026-08-19); models not listed fall back to the vendor's official API price (permanently discounted tiers at the rate users actually pay); when neither exists, the closest Artificial Analysis-scored model proxies the price and is flagged estimated. Credit-based plans (GLM / MiMo) convert credits to tokens with the vendors' official multiplier tables; peak/off-peak tiers quote the full-peak figure with off-peak as the upper bound. Quotas marked ~ are the vendor's rough figures. Claude Equiv. is anchored on measured full-use subscription data (Max 20x ≈ $3,900/mo API-equivalent at list prices), not an average-user baseline. The vs Claude column scales every plan's Equiv. against Claude Pro's $195.
Equiv. API $ = monthly allowance in turns × official API price of the flagship model
Value × = Equiv. API $ / monthly list price
When a vendor publishes only a ratio or an approximate task count (Kimi “About N Agent uses”), we infer Equiv. API $ from those official figures and mark the cell inferred. The ⓘ bubble prints the arithmetic. We still do not invent a token count.
Inferred rows: Kimi Equiv. = official About-N Agent uses × measured ¥3.6/use (one ¥199 Allegretto cycle realized ¥540 API value across 24 coding sessions, cap never hit; official Extra Usage consumer-task example ¥1.6 is the floor) ÷ FX; other Kimi tiers extrapolate linearly by use count; Claude Equiv. = measured Max 20x full-use anchor (≈$3,900/mo subscription telemetry at list prices) scaled by the official tier ratio, turns derived via the Sonnet 5 turn; Cursor Equiv. = official Other Models pool + first-party+Auto reversed from a 2026-08 Ultra dashboard (exclude BYOK), then scaled by the official 1×/3×/20× agent ratios. Official included pools (e.g. Zed's $5 hosted tokens) stay quoted at face value in the ⓘ bubble; Equiv. floors to the monthly seat so Value× ≥ 1. Qoder Equiv. uses the official credit-pack rate ($20 / 1,500 Intl; ¥40 / 1,000 CN) rather than fee-parity; CN and Intl credits are not equivalent. Measured or inferred full-use figures (Claude, Kimi, SuperGrok) are not vendor-published amounts.
Regional pricing: CN plans with a separate CNY pricebook show both tracks on the detail page. Chinese pages display the CN track as primary and recalculate Value × / per-1M costs with a manual FX snapshot (7.1 CNY/USD, 2026-08-20); all rankings on non-Chinese pages stay on the USD track. Quota parity between tracks is assumed and flagged where unverified.
GLM current-plan credits: peak 1.0× / off-peak 0.5×, Monday–Friday 14:00–18:00 UTC+8, identical on docs.bigmodel.cn and docs.z.ai. Detail pages show a workday-effective rate from overlap with 09–18 and a same-clock workday value × (displayed value × ÷ eff). Rankings, hero, and the catalog stay on the full-peak figure. A China workday overlaps peak by 4h (0.72×); a US-East workday overlaps 0h (0.50×). The retired V2 personal plan billed 3× at peak and is not used here.
Quotas and parallel sessions
Quota fields stay split: token / credits / prompts / requests, and 5-hour / weekly / monthly windows. Parallel sessions are official when published, otherwise —. Community measurements are a separate column. Capacity / $ prefers official monthly tokens; if a vendor only publishes credits or a ratio, the chart uses tokens implied from Equiv. API $.
History
Each plan is a time series: list price, yearly effective, quota fields, parallel, models, and status. A sparkline on the leaderboard is the last 90 days. The plan page has the full axis. Every production snapshot must keep captured_at, source_url, and official | estimated | unknown.
Promo resets
A promo reset is a vendor-initiated refill outside the published 5-hour / weekly / monthly window — outage compensation, a milestone gift, or a temporary multiplier. The Reset chip appears for 14 days after the last event (UTC calendar day) or while a published bonus is still valid. Codex history is seeded from the community tracker codex-resets.com (a secondary source), not an official API. Promo resets do not change Equiv. API $ or Value ×.
This build
The public site is static HTML plus one shared, cacheable plans.json. data/catalog.py is the source. scripts/build.py renders one folder per language into dist/. Filters, charts, column toggles, and search run in the browser after a single catalog fetch.