What this is worth testing
Your agent keeps the larger task and any action that follows. System One asks Jev for a typed answer on this check.
- Research value
- LowEvidence confidence is low. This rating describes the research record, not a live Jev probability or a measured saving.
- Sample size
- About 43 input tokensThe recipe sample length divided by four. This is a planning estimate, not a measured tokenizer count.
$0.02 at Jev's $0.042 per million list rate
$1.29 at a $3 per million example rate
$1.27 input-cost difference for 10,000 uses of this sample
Excludes output charges, host tool calls, retries, hosting, and paid plans. Equal input volume does not establish equal answer quality or savings on a fixed-price subscription. $3 per million is an example rate, not a quoted model price.
When an agent uses this
System One asks Jev for a typed answer. Your agent keeps the larger task, the original evidence, and any action that follows.
Example evidence
This is the recipe sample. Replace it with the evidence from your task. It is not a measured result.
{
"records": [
{
"id": "health",
"body": "GET /health returned 200 in 12 ms",
"severityNumber": 9
},
{
"id": "failure",
"body": "Database write failed: disk full",
"severityNumber": 17
}
]
}What your agent keeps
Preserve every original record. Errors, protected events and uncertainty stay available for review.
Finding
The available first-party study does not demonstrate net savings.
Research value: low. Evidence confidence: low. That rating describes the research record, not a live Jev probability.
Try it
Connect System One, then ask your agent for recipe log-triage. Inspect the input on the recipe page before you send your own evidence.