Jev vs Claude Haiku 4.5: cost per decision
Haiku is the usual baseline for high-volume classification and routing. LiteLLM measured Jev classifying 5.43x faster at 96% lower cost on their router corpus.
| Scenario | Jev 1.13 | Claude Haiku 4.5 | Ratio |
|---|---|---|---|
Light 10k items/day, 300-token state, 2 questions | $5.29 | $171.00 | 32x |
Production 100k items/day, 500-token state, 4 questions | $93.24 | $2,970 | 32x |
High volume 2M items/day, 300-token state, 1 question | $856.80 | $23,400 | 27x |
Monthly, 30 days. Both sides answer all questions for an item in a single call. Claude Haiku 4.5 additionally pays for its output tokens at $5/M; Jev 1.13 has no output charge. List prices at standard tier, excluding caching and batch discounts.
Where Claude Haiku 4.5 still wins
Haiku handles open-ended extraction where you cannot enumerate the options in advance. Jev needs the answer space defined up front.
The honest caveat
This compares price, not accuracy. A model that is 100x cheaper and wrong 5% more often can easily be the more expensive choice once you price the mistakes. Jev returns a calibrated probability with every answer, so the practical move is to route the confident cases automatically and send the rest to Claude Haiku 4.5 or a human — which makes this a routing decision rather than a replacement decision.