Track token usage across AI coding agents from your terminal. Global leaderboard with trillions of tokens tracked.
Joined February 2026
- Tweets81
- Following53
- Followers64
- Likes108
Pinned Tweet
Tokscale. For the blue hour.
You opened your laptop for one small fix. Now the sky outside is turning blue.
Photographers set alarms for the blue hour. You just never went to bed. The French have an older name for it: l’heure entre chien et loup, the hour between dog and wolf. The phrase is more than seven hundred years old. It names the moment when your own dog and a wolf start to look the same. (So do a good diff and a bad one.)
If you code with agents, you know how these nights go. You try one approach, then another. You switch models after the first one fixes a failing test by deleting it. You rerun a prompt you were sure would work the first time. By morning, the commit keeps only the part that worked.
@Tokscale counts all of it. It’s an open-source tool that tracks your token usage and what it would cost, model by model, across more than 50 AI clients. Claude Code, Codex, Cursor, and OpenCode are on the list. So are Hermes Agent, Prime Agent, and Aside. You can turn your usage into a public profile and see where you stand on the global leaderboard.
Our name borrows from the Kardashev scale, which ranks civilizations by how much energy they can harness. For anyone building with agents, that energy is tokens.
Run npx tokscale@latest to see your own numbers. Next time the sky turns blue, you’ll know what the night cost.
🚨 BREAKING: SK HYNIX HALTS SOLIDIGM'S U.S. IPO PLANS.
SK hynix has halted plans to list its NAND flash subsidiary Solidigm in the U.S., a deal Reuters reported could have valued the business at up to $150 billion and raised as much as $15 billion.
Solidigm was formed after SK hynix acquired Intel's NAND and SSD business for roughly $9 billion in 2020. The proposed IPO was intended to help finance a new U.S. semiconductor fab as AI-driven demand for enterprise SSDs continues to grow.
The plan faced strong opposition from existing shareholders concerned that a separate listing would dilute their economic exposure to Solidigm and deepen Korea's longstanding double-listing discount.
South Korean regulators also questioned whether the IPO would violate recently tightened rules protecting minority shareholders.
SK hynix officially cited market volatility and declining semiconductor share prices as reasons for discontinuing the IPO and pre-IPO financing review.
Separately, Reuters reported that SK hynix is seeking U.S. approval to install American chipmaking equipment at its Dalian, China facility, with negotiations reportedly linked to plans for a new U.S. fab. No direct connection to the IPO cancellation has been established.
Solidigm's expansion plans remain intact. SK hynix says it will continue evaluating alternative financing options, leaving the central capital allocation question unresolved: how to fund additional NAND capacity without compromising existing shareholder value.
27.5B tokens for $417.
A different kind of leaderboard flex.
Who else is optimizing for cost per token?
$417 of inference last month. 27.5B tokens. Almost certainly the #1 DeepSeek user on @tokscale.
Nobody above me on the board is close to this cheap per token. Most people buy the most capable model. A few quietly run the most efficient one into the ground.
@deepseek_ai
Your sessions, in detail.
Open the Sessions tab and press Enter to inspect a session's workspace, models, token breakdown, and estimated cost.
Available now in Tokscale v4.19.0.
we did not program this attitude into tokscale 😭
For some reason I’m going to miss this pattern when it ages out of my tokscale history
#swarmlife #doyouevenswarmbro
Vincent has crossed 1 trillion tokens over the last 12 months. Officially a token trillionaire.
He also raises a fair point about the cost figure. Recorded token usage and estimated API cost are different things. The dollar estimate isn’t necessarily what someone paid, especially when they’re accessing models through subscriptions. It should be read as a reference, not an actual bill.
What I find interesting is having a full year of usage to look back on. Individual sessions add up, and the records can get scattered across different coding agents. Bringing them together makes it easier to understand how your usage has changed over time. Tokscale’s project-level breakdown adds another view: which codebases that usage went into.
Thanks for sharing the milestone, Vincent. Seeing people measure a year of their agent usage in trillions has been a pretty surreal part of building Tokscale.
Jeffrey has reached 3 trillion tokens tracked on Tokscale, and says the actual total is much higher.
In September alone, he recorded 842 billion tokens with an estimated API list-price cost of $409,000. He reports paying about 3.4% of that amount through subscriptions.
That distinction matters: the API cost estimate and the subscription bill describe different things. The first puts a price on the recorded usage at API rates. The second is what he says he actually paid to access those tools.
For people running multiple coding agents, keeping track of the combined usage is a problem of its own. Tokscale brings those records together so you can review the total without piecing it together from each tool separately.
.@channprj has recorded 92.1 billion tokens on Tokscale and is considering moving some workloads to local models. The concern now is the cost of buying the hardware.
Tokscale 에 기록을 시작한 이후로 총 921억 토큰을 사용했다. 일부 작업을 로컬 장비의 LLM 사용으로 돌려야 할 것 같은데 요새 하드웨어가 너무 비싸서 걱정이다. 🥲 하 진작 살껄... (...라고 또 몇달 뒤에 같은 말 반복할 사람)
tokscale.ai/u/channprj
Tokscale retweeted
lol, just updated my tokscale submissions across my fleet of dev machines and now it's over 2T tokens. This also doesn't include a bunch from before February of 2026. My top day was 43.8B tokens but there are a bunch of days close to that high. This is about to explode even more
Tokscale retweeted
Nice, I hit 3 trillion tokens (tracked on tokscale, that is; the true number is much higher). The month of September I used 842 billion tokens with a list-price API cost of $409k. Glad I can pay 3.4% of that amount and get the same thing via subscriptions. Enjoy it while you can!
Tokscale retweeted
Tokscale 에 기록을 시작한 이후로 총 921억 토큰을 사용했다. 일부 작업을 로컬 장비의 LLM 사용으로 돌려야 할 것 같은데 요새 하드웨어가 너무 비싸서 걱정이다. 🥲 하 진작 살껄... (...라고 또 몇달 뒤에 같은 말 반복할 사람)
tokscale.ai/u/channprj
Tokscale retweeted
I'm officially a token trillionaire. Been tracking my token burn and surpassed 1T tokens burn in the last 12 months. Sus about the cost but burn is very real!
Tokscale retweeted
ok i spent $1.3M worth of tokens building this big project OmO alone (69.5K stars, 4M+ downloads)
according to @_junhoyeo 's tokscale.ai
