Analytics

Cost and performance across every agent on this machine · last 30 days

Total spend
$372 est.
list-price estimate, last 30 days
Sessions
211
Avg cost / session
$4.69
Tokens
59.8M
in + out, cache reads included
Failure rate
1.9%
4 of 211 finished sessions
Tool latency p50 / p95
4.48s / 8.66s

Daily spend (est.)

$19.21$38.42
Jun 4Jun 11Jun 18Jun 25Jul 3

Spend by model (est.)

claude-fable-5$207 · 29 sessions
claude-opus-4-8$80.18 · 22 sessions
claude-sonnet-5$7.68 · 6 sessions
claude-haiku-4-5$3.22 · 6 sessions

Every spend figure on this page is an estimate: recorded token counts priced at published list rates, which is not what a Max or Pro subscription actually bills. It covers the last 30 days only — not all time — and at most the newest 64 sessions, each including the subagents it spawned. The rates are in Settings, edited via pricing.json.

Tool latency

Read
0.18s · 0.9s
Grep
0.22s · 1.1s
Edit
0.35s · 1.6s
Bash
2.1s · 14.8s
Write
0.40s · 1.9s
WebFetch
1.8s · 7.2s
0.1s1s10s
p50 · p95

Wall-clock time per tool call, log scale — top 10 tools by volume. Tools whose p95 exceeds 300s are left out: they were waiting on you, not on a model.

Tokens per day

Jun 4Jul 3

Sessions per day

Jun 4Jul 3