Orsik
@orsikSince Oct 5, 2026: 277M tokens, 194 h, 317 sessions, $4,267 at API prices (est.)
Every synced day
Agents
95%Claude Code, 188 h, 263M tokens
5%Codex, 10 h, 13.6M tokens
Details
12MInput tokens
40.2MOutput tokens
224MCache writes
12BCache reads, not in the headline
256 hAgent time, parallel sessions each count
9Most sessions at once
1,869Prompts
20 daysLongest streak
Automated runs, not in hours: 74 sessions, 10 h of their own wall-clock time, 3 h of it outside your interactive sessions. These are claude -p, Agent SDK and codex exec runs.
| Model | Tokens | Share | At API prices (est.) |
|---|---|---|---|
| claude-opus-5-5 | 176M | 64% | $3,237 |
| claude-sonnet-5-5 | 74.5M | 27% | $771 |
| gpt-6.1-sol | 12.7M | 5% | $159 |
| claude-haiku-5-5 | 9.39M | 3% | $5.47 |
| claude-fable-5-1 | 2.56M | 1% | $71 |
| gpt-6-astra | 763K | <1% | $21 |
| claude-haiku-4-5-20251001 | 274K | <1% | $0.76 |
| gpt-6.1-terra | 150K | <1% | $0.27 |
Make your own
One command. It reads your coding agent logs on your computer and makes this card.
$ npx codewrapped