This is a real 30-day Anthropic workload (Claude Code + SDK), semi-anonymized — totals rounded, customer IDs stripped. Connect your own API key to see your numbers in the same shape.
Sample AI spending overview — 30-day window
+100.0% vs last period
Across 2 providers
$4655.20 (92%)
Trend: rising ↑
| Provider | ||||
|---|---|---|---|---|
| claude-sonnet-4-5 | Anthropic | 14.8K | $2700.02 | 53.36% |
| claude-opus-4-7 | Anthropic | 1.2K | $1443.11 | 28.52% |
| claude-haiku-4-5 | Anthropic | 38.1K | $512.07 | 10.12% |
| gpt-4o | OpenAI | 980 | $283.36 | 5.6% |
| gpt-4o-mini | OpenAI | 6.4K | $121.44 | 2.4% |
A meaningful slice of Opus traffic looks like routine classification and short-form generation — exactly the workload Sonnet 4.5 handles at ~3.3x lower cost with no measurable quality drop on this owner's benchmark set.
You have 1 more suggestions. Upgrade to Pro to see all optimization opportunities.
Unlock All Suggestions30 days of Anthropic-heavy spend with an OpenAI sliver.
Daily chart, model breakdown, and optimization suggestion are the same components the dashboard renders for paying users.
Start FreeSee which customers cost more than they pay — without a proxy.
| Customer | Revenue | AI cost | Margin |
|---|---|---|---|
Globex | $499.00 | $212.00 | $287.00 |
Initech | $299.00 | $168.00 | $131.00 |
Hooli | $199.00 | $154.00 | $45.00 |
Acme CorpUnprofitable | $99.00 | $140.00 | $-41.00 |
SoylentUnprofitable | $49.00 | $88.00 | $-39.00 |