Image JevBench v0.3.0 · individual system
Image Hopper (HopitAI)
LLM decoderBase model: undisclosedlocal/GPU evaluation
Image JevBench v0.3.0 composite score
15.300
Rank #16 of 44 ranked systems.
From the published full-benchmark aggregate. The composite combines intelligence, calibration, speed and cost.
Published axes
- intelligence
- 32.5
- calibration
- 70.9
- speed
- 83.2
- cost
- 42.7
Cost and speed evidence
- Cost per 1,000 decisions
- USD 0.077291
- Cost basis
- measured GPU seconds x $1.19/GPU-hour
- Measured latency · p50 / p95
- 0.175 s / 0.409 s
- Adjusted latency · p50 / p95
- 0.499 s / 0.968 s
- Latency adjustment
- v1.4 local latency adjustment
- Public accuracy
- 53.6%
- Sealed accuracy
- 55.0%
Published 2026-10-06. Scores and ranks can change in a later release. See the method notes.