What these tokens would have cost billed per token at each model's published price. It is the one figure that compares across people on different plans.
List-rate estimate / day
Jun 20, 2026peak $666.28Sep 11, 2026
on a plan$5,757
paid per token$259.37
per active day$71.63
7.51B
tokens
89.4M on a working day
74,721
requests
100.5K each
84
active days
not a day missed
84
day streak
still running
Activity
+731% vs previous 30 days
7.51B tokens over 84 active days · last 12 months
Averaging 89.4M on a day worked, with a median of 31.2M. Busiest was Aug 18, 2026 at 750.3M. Cells are shaded by what the day was worth at list rates, not by token count. Hover a day for the models and harnesses behind it.
less spendmore spend
Last 12 months in full on this profile, last synced Sep 11, 2026. Counters only: no prompts, no project names, no file paths.
Drift
What got reached for, month by month
Each row is one month at 100% of its own tokens, so a quiet month reads the same width as a loud one and the only thing changing is the mix.
Output is a fraction of the tokens and usually most of the money, because it is priced at several times the rate of anything read back out of cache.
cache write
$2,954
cache read
$2,057
output
$872.54
fresh input
$132.57
Every model in this archive has a known rate.
Cache
Reuse, not volume
A raw token total mostly measures how much context got re-read. The ratios underneath it are the part that says anything about how the work was done.
97.1%
of everything entering the context window came from cache instead of being paid for again
97.6×
tokens of context read per token written
100.5K
tokens in an average request
cache read
89% · 6.71B
cache write
7% · 518.2M
output
1% · 76.2M
fresh input
3% · 202.8M
Models
33 models, mostly claude-sonnet-5
claude-sonnet-5
2.68B · 35.7%
$1,336 estimated at list rates · 21,325 requests
claude-opus-5
2.37B · 31.5%
$2,862 estimated at list rates · 15,614 requests
muse-spark-1.3
1.04B · 13.8%
$258.12 estimated at list rates · 12,374 requests
muse-spark-1.3-contributor-free
283.9M · 3.8%
$13.16 estimated at list rates · 5,512 requests
claude-fable-5-1
224.0M · 3%
$397.39 estimated at list rates · 759 requests
claude-fable-5
185.0M · 2.5%
$585.33 estimated at list rates · 2,124 requests
gemini-3.8-flash
177.2M · 2.4%
$24.03 estimated at list rates · 1,084 requests
claude-opus-4-8
131.5M · 1.8%
$220.81 estimated at list rates · 1,448 requests
claude-haiku-4-5
123.3M · 1.6%
$66.23 estimated at list rates · 4,240 requests
claude-sonnet-4-5
89.0M · 1.2%
$117.81 estimated at list rates · 4,349 requests
+23 more below one percent.
Tools
5 harnesses, one archive
Claude Code
5.84B · 52,046 requests · $5,673
77.8%
Muse
1.01B · 11,753 requests · $242.73
13.5%
OpenCode
344.1M · 6,660 requests · $25.96
4.6%
Antigravity
303.5M · 3,999 requests · $70.93
4%
Codex
12.5M · 263 requests · $3.89
0.2%
Rhythm
Heaviest on Thursday
Tokens by day of the week, in the timezone each device reported.
Mon: 1.19B over 12 days
Tue: 1.34B over 12 days
Wed: 758.3M over 12 days
Thu: 1.91B over 12 days
Fri: 1.19B over 12 days
Sat: 514.0M over 12 days
Sun: 608.9M over 12 days
MTWTFSS
By month
4 months tracked, busiest Aug 26.
Jun 26: 16.9M over 11 days
Jul 26: 889.7M over 31 days
Aug 26: 4.17B over 31 days
Sep 26: 2.43B over 11 days
JunJulAugSep
Hour of day is not in the sync payload yet, so the rhythm here stops at the day. See docs/tokenstat-sync-cli.md for what the CLI would need to send.
Spread
Across 1791 projects
Project names stay on the device. What reaches us is an opaque hash we can count but never read, so a profile can show how the work was spread and never which work it was.
project 1
43.7%
project 2
23.2%
project 3
20.3%
project 4
6.5%
project 5
1.8%
project 6
1.3%
project 7
1.1%
project 8
1%
Ordered by size, unlabelled on purpose. The top project is 43.7% of everything.
How the counts were read
Some harnesses report token counts exactly. Others have to be derived from what they write to disk. The archive keeps that distinction rather than flattening it, so a number that was inferred never passes itself off as a measurement.
