Image JevBench v0.1.5 · individual system
GPT-6 Luna (low reasoning effort)
Closed APIBase model: undisclosedArchived v0.1.5 score, not ranked on the v0.3 scale. See the v0.1.5 archive.
the operator's hosted API
Image JevBench v0.1.5 composite score
16.056
v0.1.5 rank #44 of 50 (archived).
From the published full-benchmark aggregate. The composite combines intelligence, calibration, speed and cost.
Published axes
- intelligence
- 81.1
- calibration
- 88.3
- speed
- 65.7
- cost
- 27.5
Cost and speed evidence
- Cost per 1,000 decisions
- USD 0.216570
- Cost basis
- OpenRouter usage.cost receipts
- Measured latency · p50 / p95
- 2.905 s / 9.256 s
- Adjusted latency · p50 / p95
- 2.905 s / 9.256 s
- Latency adjustment
- none (hosted API, measured as is)
- Public accuracy
- 89.0%
- Sealed accuracy
- 86.6%
Published 2026-10-01. Scores and ranks can change in a later release. See the method notes.