Image JevBench v0.3.0 · individual system
Clef (Cloudflare)
LLM decoderBase model: undisclosedlocal/GPU evaluation
Image JevBench v0.3.0 composite score
18.652
Rank #13 of 44 ranked systems.
From the published full-benchmark aggregate. The composite combines intelligence, calibration, speed and cost.
Published axes
- intelligence
- 47.4
- calibration
- 68.8
- speed
- 75.7
- cost
- 32.2
Cost and speed evidence
- Cost per 1,000 decisions
- USD 0.182067
- Cost basis
- measured GPU seconds x $1.19/GPU-hour
- Measured latency · p50 / p95
- 0.515 s / 1.057 s
- Adjusted latency · p50 / p95
- 1.180 s / 2.265 s
- Latency adjustment
- v1.4 local latency adjustment
- Public accuracy
- 62.6%
- Sealed accuracy
- 65.6%
Published 2026-10-06. Scores and ranks can change in a later release. See the method notes.