AGENTWACHMonthly usage report
Acme Robotics — Production
Mar 1, 2026 – Mar 31, 2026 · Workspace: acme-prod
Report ID
rpt_20260401_acme
Generated Apr 1, 2026
Total cost
$447.91
across all providers
Agent runs
5,468
5 active agents
Input tokens
14,128,300
billable
Output tokens
5,699,450
billable
Cost by agent
| Agent | Type | Runs | Input tok | Output tok | Models | Cost |
|---|---|---|---|---|---|---|
| Lead-Qualifier | langgraph | 1,284 | 4,120,300 | 1,840,950 | gpt-4o-mini + claude-3.5-sonnet | $142.18 |
| Outbound-Writer | langchain | 612 | 1,905,000 | 2,140,700 | claude-3.5-sonnet | $187.43 |
| Inbox-Triage | openai-assistant | 3,210 | 6,842,100 | 1,204,800 | gpt-4o-mini | $64.31 |
| Report-Drafter | langgraph | 88 | 942,500 | 410,200 | gpt-4o | $41.92 |
| Web-Researcher | browser | 274 | 318,400 | 102,800 | gemini-1.5-pro | $12.07 |
| Total | 5,468 | 14,128,300 | 5,699,450 | $447.91 |
Cost by model
- claude-3.5-sonnet$198.62 (44%)1,840 calls
- gpt-4o-mini$158.04 (35%)4,112 calls
- gpt-4o$41.92 (9%)88 calls
- gemini-1.5-pro$12.07 (3%)274 calls
Notable incidents
- Mar 09·Outbound-Writer
Loop on a stale lead source — caught by budget alarm at $14.20, paused automatically.
- Mar 17·Inbox-Triage
Provider 5xx burst from OpenAI — auto-retries succeeded, no human action needed.
- Mar 24·Report-Drafter
Long-context run consumed 38k tokens — flagged for prompt review.
Incidents are auto-surfaced from budget alarms, provider 5xx clusters, and outlier runs.