Image JevBench v0.3.0 · individual system
OmniJev 0.8B (tinnel123)
adapter · 0.8BBase model: Qwen/Qwen3.5-0.8Bsourcelocal/GPU evaluation
Image JevBench v0.3.0 composite score
1.571
Rank #38 of 44 ranked systems.
From the published full-benchmark aggregate. The composite combines intelligence, calibration, speed and cost.
Published axes
- intelligence
- 11.4
- calibration
- 65.8
- speed
- 88.4
- cost
- 54.1
Cost and speed evidence
- Cost per 1,000 decisions
- USD 0.033878
- Cost basis
- measured GPU seconds x $1.19/GPU-hour
- Measured latency · p50 / p95
- 0.100 s / 0.133 s
- Adjusted latency · p50 / p95
- 0.351 s / 0.417 s
- Latency adjustment
- v1.4 local latency adjustment
- Public accuracy
- 41.7%
- Sealed accuracy
- 39.5%
Published 2026-10-06. Scores and ranks can change in a later release. See the method notes.