Skip to content
BetaBenchAlert v0.1 is in Beta.Numbers are real, but pages and rules can still change.See what changed
BenchAlertby

Quota benchmark

Quota

What 100% of a weekly limit is worth at the vendor's own API prices.

QuotaAPI $ / week

Liveupdated 17h agonext run 03:10 UTC

  1. Claude Max 20x≥ lower bound≥ $1,571

    5-hour $2.20 · Fable 5 up to 50% of weekly

  2. SuperGrok Heavy≥ lower bound≥ $1,201

    Includes Cursor Ultra — $400/mo usage on top

  3. ChatGPT Pro ($100, 5×)≥ lower bound≥ $604

    Vendor's own figure $600 (15,000 credits)

Last reading
Aug 23
Method
v1
Interval
calibrated
How we measure →
Details

Everything behind the three figures: the daily series, the cross-checks, what changed, and what this is measured on.

Read this before you quote a number

Measured on one account and one machine. Our local logs see only this PC, so Q is a lower bound — a drop can be our own blind spot, not a vendor change. The detector's thresholds are set for about one false alarm per provider per year.

Q is the dollar value of one full usage window, priced as if the same tokens had been bought at the vendor's own published API list price. The point is the line, not the level. If a vendor quietly reduces what 100% of a window buys, Q falls — and you do not have to take anyone's word for it.

Every plan in full

Two independent estimators run beside the published figure. They are not graded and are never the answer — they are here so you can see whether the methods agree.

Claude Code$200/mo, paid by us

Claude Max 20x

at least $1,571

100% of the weekly window is ≥ $1,571 at official API list price — insufficient gauge movement for a point estimate yet

Grade C · Thinlower bound

Indicative cross-checks (not graded): OLS ≈ $70.33 (R² 0.06) · floor ≈ $1,571 ($1,545 – $1,571)

The other windows on this plan

5-hour session$2.20B
Weekly — Fable 5 sub-cap (50% of weekly) · 50% of weekly≥ $879C

As of Aug 23 · window resets Thu 04:00 UTCJob ran 17h ago

Codex$100/mo, paid by us

ChatGPT Pro ($100, 5×)

at least $604

100% of the weekly window is ≥ $604 at official API list price — insufficient gauge movement for a point estimate yet

Grade — · Gapslower bound

Indicative cross-checks (not graded): OLS ≈ $594 (R² 1.00) · floor ≈ $604 ($592 – $604)

Vendor's own figure

$600 — 15,000 credits × $0.04 = $600 a week

Our measured value sits within a few percent of 15,000 credits = $600.

As of Aug 23 · window resets Tue 00:30 UTCJob ran 17h ago

Grok Build$300/mo (as billed)

SuperGrok Heavy

at least $1,201

100% of the weekly window is ≥ $1,201 at official API list price — insufficient gauge movement for a point estimate yet

Grade — · Gapslower bound

Indicative cross-checks (not graded): OLS ≈ $1,161 (R² 0.99) · floor ≈ $1,201 ($1,135 – $1,201)

In Grok's own pool price

$209 a week in pool dollars

Grok's own per-run cost is an internal basis (0.17× list for grok-4.6, 0.34× for grok-4.5); we show list dollars. Nobody publishes a pool size for this plan — the figure above is our own measurement.

Comes with

  • Includes Cursor Ultra — Cursor's own promotion for Heavy subscribers (claimed inside Cursor; free while Heavy renews): normally $200/mo, includes ≥ $400/mo (≈ $92/week) of API-priced model usage per cursor.com/docs/account/pricing

As of Aug 23 · window resets Sun 17:53 UTCJob ran 17h ago

Claude Max 20x over time

Claude Code · Weekly (all models). One point per day, newest on the right.

Weekly (all models)

100% of the weekly window, day by day

Every point is one day's answer, in the vendor's own published API list dollars.

Published numbercalibrated intervallower bound only (≥)flagged day
list $0$625$1,250$1,875$2,500
Jun 25Jul 1Jul 7Jul 13Jul 19Jul 25Jul 31Aug 6Aug 12Aug 18Aug 23

How to read this

Each point is one day's answer to the same question: what would 100% of this window cost at the vendor's own API price? The shaded ribbon is the calibrated interval — not a confidence interval, but an envelope checked against simulated data where the true answer was known. An open marker over a shaded column is a lower bound: the gauge did not move enough that day for a point estimate, so the plot shows the floor and draws no band above it. A red mark on the rail underneath is a day we flagged — the table names it.

Full record — every day, every number
Day100% of windowCalibrated intervalGradeClean stretchesMeter at day endFlags
Aug 23≥ $1,571OLS ≈ $70.33 (R² 0.06) · floor ≈ $1,571 ($1,545 – $1,571)—C · Thin659%Meter barely moved
Aug 22≥ $2,154OLS ≈ $1,020 (R² 0.48) · floor ≈ $2,154 ($2,105 – $2,154)—— · Gaps243%Fell back · Missed readings
Aug 21≥ $2,374OLS ≈ $1,939 (R² 0.72) · floor ≈ $2,374 ($2,314 – $2,374)—— · Not enough yet039%Fell back
Aug 20——— · Not enough yet0—Fell back
Aug 19——— · Not enough yet0—Fell back
Aug 18——— · Not enough yet0—Fell back
Aug 17——— · Not enough yet0—Fell back
Aug 16——— · Not enough yet0—Fell back
Aug 15——— · Not enough yet0—Fell back
Aug 14——— · Not enough yet0—Fell back
Aug 13——— · Not enough yet00%Fell back
Aug 12——— · Not enough yet0—Fell back
Aug 11——— · Not enough yet0—Fell back
Aug 10——— · Not enough yet0—Fell back
Aug 9——— · Not enough yet0—Fell back
Aug 8——— · Not enough yet0—Fell back
Aug 7——— · Not enough yet0—Fell back
Aug 6——— · Not enough yet0—Fell back
Aug 5——— · Not enough yet0—Fell back
Aug 4——— · Not enough yet0—Fell back
Aug 3——— · Not enough yet0—Fell back
Aug 2——— · Not enough yet0—Fell back
Aug 1——— · Not enough yet0—Fell back
Jul 31——— · Not enough yet0—Fell back
Jul 30——— · Not enough yet0—Fell back
Jul 29——— · Not enough yet0—Fell back
Jul 28——— · Not enough yet0—Fell back
Jul 27——— · Not enough yet0—Fell back
Jul 26——— · Not enough yet0—Fell back
Jul 25——— · Not enough yet0—Fell back
Jul 24——— · Not enough yet0—Fell back
Jul 23——— · Not enough yet0—Fell back
Jul 22——— · Not enough yet0—Fell back
Jul 21——— · Not enough yet0—Fell back
Jul 20——— · Not enough yet0—Fell back
Jul 19——— · Not enough yet0—Fell back
Jul 18——— · Not enough yet0—Fell back
Jul 17——— · Not enough yet0—Fell back
Jul 16——— · Not enough yet0—Fell back
Jul 15——— · Not enough yet0—Fell back
Jul 14——— · Not enough yet0—Fell back
Jul 13——— · Not enough yet0—Fell back
Jul 12——— · Not enough yet0—Fell back
Jul 11——— · Not enough yet0—Fell back
Jul 10——— · Not enough yet0—Fell back
Jul 9——— · Not enough yet0—Fell back
Jul 8——— · Not enough yet0—Fell back
Jul 7——— · Not enough yet0—Fell back
Jul 6——— · Not enough yet0—Fell back
Jul 5——— · Not enough yet0—Fell back
Jul 4——— · Not enough yet0—Fell back
Jul 3——— · Not enough yet0—Fell back
Jul 2——— · Not enough yet0—Fell back
Jul 1——— · Not enough yet0—Fell back
Jun 30——— · Not enough yet0—Fell back
Jun 29——— · Not enough yet0—Fell back
Jun 28——— · Not enough yet0—Fell back
Jun 27——— · Not enough yet0—Fell back
Jun 26——— · Not enough yet0—Fell back
Jun 25——— · Not enough yet0—Fell back
60 days shown · same rows as /api/quota?provider=claudePrice table 2026-08-21

Claude Code · Weekly (all models)

Meter readings today
326
Clean stretches used
6
Grade
C · Thin
Last reading
17h ago

Estimator: envelope-p97. Price table: 2026-08-21. Not measured for this window yet.

ChatGPT Pro ($100, 5×) over time

Codex · Weekly. One point per day, newest on the right.

Weekly

100% of the weekly window, day by day

Every point is one day's answer, in the vendor's own published API list dollars.

Published numbercalibrated intervallower bound only (≥)flagged day
list $0$250$500$750$1,000
Jun 25Jul 1Jul 7Jul 13Jul 19Jul 25Jul 31Aug 6Aug 12Aug 18Aug 23

How to read this

Each point is one day's answer to the same question: what would 100% of this window cost at the vendor's own API price? The shaded ribbon is the calibrated interval — not a confidence interval, but an envelope checked against simulated data where the true answer was known. An open marker over a shaded column is a lower bound: the gauge did not move enough that day for a point estimate, so the plot shows the floor and draws no band above it. A red mark on the rail underneath is a day we flagged — the table names it.

Full record — every day, every number
Day100% of windowCalibrated intervalGradeClean stretchesMeter at day endFlags
Aug 23≥ $604OLS ≈ $594 (R² 1.00) · floor ≈ $604 ($592 – $604)—— · Gaps1649%Missed readings
Aug 22≥ $588OLS ≈ $582 (R² 1.00) · floor ≈ $588 ($574 – $588)—— · Gaps1340%Missed readings
Aug 21$649OLS ≈ $582 (R² 1.00) · floor ≈ $599 ($584 – $599)$569 – $807B · Good1239%—
Aug 20$649OLS ≈ $582 (R² 1.00) · floor ≈ $596 ($581 – $596)$569 – $807B · Good1238%—
Aug 19$649OLS ≈ $581 (R² 1.00) · floor ≈ $581 ($566 – $581)$569 – $807B · Good1238%—
Aug 18$641OLS ≈ $581 (R² 1.00) · floor ≈ $588 ($568 – $588)$561 – $796B · Good929%Plan changed · Speed tier changed
Aug 17$163OLS ≈ $126 (R² 0.90) · floor ≈ $127 ($110 – $139)$143 – $203A · Solid69—Reset time moved · Meter went backwards
Aug 16$163OLS ≈ $126 (R² 0.90) · floor ≈ $127 ($110 – $139)$143 – $203A · Solid6986%Reset time moved · Meter went backwards
Aug 15$163OLS ≈ $126 (R² 0.90) · floor ≈ $127 ($110 – $139)$143 – $203A · Solid6986%Reset time moved · Meter went backwards
Aug 14$163OLS ≈ $126 (R² 0.89) · floor ≈ $127 ($110 – $139)$143 – $203A · Solid6884%Reset time moved · Meter went backwards
Aug 13$163OLS ≈ $127 (R² 0.98) · floor ≈ $127 ($110 – $139)$143 – $202A · Solid4124%Reset time moved · Meter went backwards
Aug 12$168OLS ≈ $127 (R² 1.00) · floor ≈ $126 ($110 – $139)$148 – $209A · Solid3918%Reset time moved
Aug 11$168OLS ≈ $124 (R² 0.99) · floor ≈ $127 ($110 – $130)$148 – $209B · Good170%Reset time moved
Aug 10$168OLS ≈ $124 (R² 0.99) · floor ≈ $127 ($110 – $130)$148 – $209B · Good1731%Reset time moved
Aug 9$168OLS ≈ $121 (R² 0.99) · floor ≈ $126 ($110 – $130)$148 – $209B · Good1628%Reset time moved
Aug 8$156OLS ≈ $120 (R² 0.99) · floor ≈ $121 ($108 – $130)$137 – $194B · Good1011%Reset time moved
Aug 7≥ $130OLS ≈ $134 (R² 0.98) · floor ≈ $130 ($120 – $130)—C · Thin4—Meter barely moved · Reset time moved
Aug 6≥ $130OLS ≈ $134 (R² 0.98) · floor ≈ $130 ($120 – $130)—C · Thin4—Meter barely moved · Reset time moved
Aug 5≥ $130OLS ≈ $134 (R² 0.98) · floor ≈ $130 ($120 – $130)—C · Thin4—Meter barely moved · Reset time moved
Aug 4≥ $130OLS ≈ $134 (R² 0.98) · floor ≈ $130 ($120 – $130)—C · Thin44.0%Meter barely moved · Reset time moved
Aug 3≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved · Reset time moved
Aug 2≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved · Reset time moved
Aug 1≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved · Reset time moved
Jul 31≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved · Reset time moved
Jul 30≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved · Reset time moved
Jul 29≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved · Reset time moved
Jul 28≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved
Jul 27≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved
Jul 26≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved
Jul 25≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved
Jul 24≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved
Jul 23≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin3—Meter barely moved
Jul 22≥ $130OLS ≈ $135 (R² 0.99) · floor ≈ $130 ($120 – $130)—C · Thin312%Meter barely moved · Window length changed
Jul 21$21.07OLS ≈ $16.20 (R² 0.94) · floor ≈ $15.46 ($15.15 – $18.95)$18.46 – $26.19B · Good14—Reset time moved
Jul 20$21.07OLS ≈ $16.20 (R² 0.94) · floor ≈ $15.46 ($15.15 – $18.95)$18.46 – $26.19B · Good14—Reset time moved
Jul 19$21.07OLS ≈ $16.20 (R² 0.94) · floor ≈ $15.46 ($15.15 – $18.95)$18.46 – $26.19B · Good14—Reset time moved
Jul 18$21.07OLS ≈ $16.20 (R² 0.94) · floor ≈ $15.46 ($15.15 – $18.95)$18.46 – $26.19B · Good14—Reset time moved
Jul 17$21.07OLS ≈ $16.20 (R² 0.94) · floor ≈ $15.46 ($15.15 – $18.95)$18.46 – $26.19B · Good14—Reset time moved
Jul 16$21.07OLS ≈ $16.20 (R² 0.94) · floor ≈ $15.46 ($15.15 – $18.95)$18.46 – $26.19B · Good14—Reset time moved
Jul 15$21.07OLS ≈ $16.20 (R² 0.94) · floor ≈ $15.46 ($15.15 – $18.95)$18.46 – $26.19B · Good14—Reset time moved
Jul 14$21.07OLS ≈ $16.20 (R² 0.94) · floor ≈ $15.46 ($15.15 – $18.95)$18.46 – $26.19B · Good14—Reset time moved
Jul 13$21.07OLS ≈ $16.20 (R² 0.94) · floor ≈ $15.46 ($15.15 – $18.95)$18.46 – $26.19B · Good14—Reset time moved
Jul 12$21.07OLS ≈ $16.20 (R² 0.94) · floor ≈ $15.46 ($15.15 – $18.95)$18.46 – $26.19B · Good1482%Quota dropped · Reset time moved
Jul 11≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jul 10≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jul 9≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jul 8≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jul 7≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jul 6≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jul 5≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jul 4≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jul 3≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jul 2≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jul 1≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jun 30≥ $18.95OLS ≈ $16.20 (R² 0.84) · floor ≈ $18.95 ($17.77 – $18.95)—— · Not enough yet2—Fell back · Reset time moved
Jun 29$64.65OLS ≈ $35.86 (R² 0.98) · floor ≈ $39.05 ($17.77 – $46.19)$56.66 – $80.39A · Solid29—Reset time moved
Jun 28$64.65OLS ≈ $35.86 (R² 0.98) · floor ≈ $39.05 ($17.77 – $46.19)$56.66 – $80.39A · Solid29—Reset time moved
Jun 27$64.65OLS ≈ $35.86 (R² 0.98) · floor ≈ $39.05 ($17.77 – $46.19)$56.66 – $80.39A · Solid29—Reset time moved
Jun 26$64.65OLS ≈ $35.86 (R² 0.98) · floor ≈ $39.05 ($17.77 – $46.19)$56.66 – $80.39A · Solid29—Reset time moved
Jun 25$64.65OLS ≈ $35.86 (R² 0.98) · floor ≈ $39.05 ($17.77 – $46.19)$56.66 – $80.39A · Solid29—Reset time moved
60 days shown · same rows as /api/quota?provider=codexPrice table 2026-08-21

Codex · Weekly

Meter readings today
7
Clean stretches used
16
Grade
— · Gaps
Last reading
17h ago

Estimator: envelope-p95. Price table: 2026-08-21. Not measured for this window yet.

SuperGrok Heavy over time

Grok Build · Weekly shared pool (Build + Chat + Imagine + Voice). One point per day, newest on the right.

Weekly shared pool (Build + Chat + Imagine + Voice)

100% of the weekly window, day by day

Every point is one day's answer, in the vendor's own published API list dollars.

Published numbercalibrated intervallower bound only (≥)flagged day
list $0$375$750$1,125$1,500
Jun 25Jul 1Jul 7Jul 13Jul 19Jul 25Jul 31Aug 6Aug 12Aug 18Aug 23

How to read this

Each point is one day's answer to the same question: what would 100% of this window cost at the vendor's own API price? The shaded ribbon is the calibrated interval — not a confidence interval, but an envelope checked against simulated data where the true answer was known. An open marker over a shaded column is a lower bound: the gauge did not move enough that day for a point estimate, so the plot shows the floor and draws no band above it. A red mark on the rail underneath is a day we flagged — the table names it.

Full record — every day, every number
Day100% of windowCalibrated intervalGradeClean stretchesMeter at day endFlags
Aug 23≥ $1,201OLS ≈ $1,161 (R² 0.99) · floor ≈ $1,201 ($1,135 – $1,201)—— · Gaps217%Fell back · Missed readings
Aug 22≥ $1,226OLS ≈ $1,121 (R² 0.97) · floor ≈ $1,226 ($1,104 – $1,226)—— · Gaps29.0%Fell back · Missed readings
Aug 21≥ $1,302floor ≈ $1,302 ($976 – $1,302)—— · Not enough yet03.0%Fell back
Aug 20≥ $426OLS ≈ $318 (R² 0.86) · floor ≈ $426 ($341 – $426)—— · Not enough yet04.0%Fell back
Aug 19——— · Not enough yet00%Fell back · Plan changed
Aug 18≥ $204OLS ≈ $295 (R² 0.95) · floor ≈ $204 ($199 – $204)—— · Not enough yet136%Fell back
Aug 17——— · Not enough yet0—Fell back
Aug 16——— · Not enough yet0—Fell back
Aug 15——— · Not enough yet0—Fell back
Aug 14——— · Not enough yet0—Fell back
Aug 13——— · Not enough yet0—Fell back
Aug 12——— · Not enough yet0—Fell back
Aug 11——— · Not enough yet0—Fell back
Aug 10——— · Not enough yet0—Fell back
Aug 9——— · Not enough yet0—Fell back
Aug 8——— · Not enough yet0—Fell back
Aug 7——— · Not enough yet0—Fell back
Aug 6——— · Not enough yet0—Fell back
Aug 5——— · Not enough yet0—Fell back
Aug 4——— · Not enough yet0—Fell back
Aug 3——— · Not enough yet0—Fell back
Aug 2——— · Not enough yet0—Fell back
Aug 1——— · Not enough yet0—Fell back
Jul 31——— · Not enough yet0—Fell back
Jul 30——— · Not enough yet0—Fell back
Jul 29——— · Not enough yet0—Fell back
Jul 28——— · Not enough yet0—Fell back
Jul 27——— · Not enough yet0—Fell back
Jul 26——— · Not enough yet0—Fell back
Jul 25——— · Not enough yet0—Fell back
Jul 24——— · Not enough yet0—Fell back
Jul 23——— · Not enough yet0—Fell back
Jul 22——— · Not enough yet0—Fell back
Jul 21——— · Not enough yet0—Fell back
Jul 20——— · Not enough yet0—Fell back
Jul 19——— · Not enough yet0—Fell back
Jul 18——— · Not enough yet0—Fell back
Jul 17——— · Not enough yet0—Fell back
Jul 16——— · Not enough yet0—Fell back
Jul 15——— · Not enough yet0—Fell back
Jul 14——— · Not enough yet0—Fell back
Jul 13——— · Not enough yet0—Fell back
Jul 12——— · Not enough yet0—Fell back
Jul 11——— · Not enough yet0—Fell back
Jul 10——— · Not enough yet0—Fell back
Jul 9——— · Not enough yet0—Fell back
Jul 8——— · Not enough yet0—Fell back
Jul 7——— · Not enough yet0—Fell back
Jul 6——— · Not enough yet0—Fell back
Jul 5——— · Not enough yet0—Fell back
Jul 4——— · Not enough yet0—Fell back
Jul 3——— · Not enough yet0—Fell back
Jul 2——— · Not enough yet0—Fell back
Jul 1——— · Not enough yet0—Fell back
Jun 30——— · Not enough yet0—Fell back
Jun 29——— · Not enough yet0—Fell back
Jun 28——— · Not enough yet0—Fell back
Jun 27——— · Not enough yet0—Fell back
Jun 26——— · Not enough yet0—Fell back
Jun 25——— · Not enough yet0—Fell back
60 days shown · same rows as /api/quota?provider=grokPrice table 2026-08-21

Grok Build · Weekly shared pool (Build + Chat + Imagine + Voice)

Meter readings today
14
Clean stretches used
2
Grade
— · Gaps
Last reading
17h ago

Estimator: envelope-p90. Price table: 2026-08-21. Not measured for this window yet.

What our own testing costs the plan

We run this lab on the same plans. This is how much of each window our own hourly checks use, so you can see the measurement is not moving the thing it measures.

Claude Max 20x

Not measured for this window yet.

ChatGPT Pro ($100, 5×)

Not measured for this window yet.

SuperGrok Heavy

Not measured for this window yet.

What changed

Every event the daily job raised in the last 30 days. A plan rename, a window that changed length, a price table update, or a gap in our own readings. We start a fresh baseline instead of reporting a rename as a shrinking limit.

  • Aug 23Claude Code

    Meter barely moved — Not enough movement that day to name a number.

  • Aug 23Claude Code

    Meter barely moved — Not enough movement that day to name a number.

  • Aug 23Codex

    Missed readings — We lost meter readings for part of the window.

  • Aug 23Grok Build

    Fell back — The main estimator refused, so a simpler one ran. The grade is capped.

  • Aug 23Grok Build

    Missed readings — We lost meter readings for part of the window.

  • Aug 22Claude Code

    Missed readings — We lost meter readings for part of the window.

  • Aug 22Claude Code

    Fell back — The main estimator refused, so a simpler one ran. The grade is capped.

  • Aug 22Claude Code

    Missed readings — We lost meter readings for part of the window.

  • Aug 22Claude Code

    Missed readings — We lost meter readings for part of the window.

  • Aug 22Codex

    Missed readings — We lost meter readings for part of the window.

  • Aug 22Grok Build

    Fell back — The main estimator refused, so a simpler one ran. The grade is capped.

  • Aug 22Grok Build

    Missed readings — We lost meter readings for part of the window.

Read next

  • How we test

    The full definition, the price basis, and the limits.

  • Logs

    Every sample we saved, and what our own testing spent.

  • Lab health

    Is the lab fresh, and is the meter still readable?

  • Developers

    The same rows as JSON at /api/quota.