Clef-Flash (Cloudflare)
LLM decoderBase model: undisclosedlocal/GPU evaluation
Image JevBench v0.3.0 composite score
20.413
Rank #12 of 44 ranked systems.
From the published full-benchmark aggregate. The composite combines intelligence, calibration, speed and cost.
Published axes
- intelligence
- 37.3
- calibration
- 67.2
- speed
- 81.4
- cost
- 42.2
Cost and speed evidence
- Cost per 1,000 decisions
- USD 0.084358
- Cost basis
- measured GPU seconds x $1.19/GPU-hour
- Measured latency · p50 / p95
- 0.234 s / 0.506 s
- Adjusted latency · p50 / p95
- 0.618 s / 1.162 s
- Latency adjustment
- v1.4 local latency adjustment
- Public accuracy
- 55.3%
- Sealed accuracy
- 59.0%
Published 2026-10-06. Scores and ranks can change in a later release. See the method notes.