JevBench v1.5.5 · individual system
Clef (Cloudflare, Qwen3.8-27B post-train with a joint schema head, multimodal, measured on text)
Jev-compatible decision model · by Cloudflare · Openness: open weights
Base model: undisclosedv1.5 roster addendum A6
apache-2.0
JevBench v1.5.5 score
16.992
Option A: rank #62 of 109 ranked systems.
The three option scores and ranks are published independently; the headline is Option A.
| Option | Score | Rank |
|---|---|---|
| A · headline | 16.992 | #62 of 109 |
| B | 17.730 | #62 of 109 |
| C | 16.992 | #59 of 109 |
Published axes
- intelligence
- 67.9
- calibration
- 86.7
- speed
- 79.9
- cost
- 28.1
Bands use the published 0–100 axis values; the marked reference is Jev 1.13.0 when that axis is available.
Run and cost evidence
- Run status
- complete · 1,624 decisions · 0 missing
- Cost
- $0.249 per 1,000 decisions · estimate · price floor: base-model reference price applied; self-hosted estimate, not measured hosted billing
- Median latency
- 0.866 seconds, adjusted · x2 + 0.15 s (assumption, not measured)
- Endpoint condition
- evaluator-owned Lium GPU pod (H100 80GB), offline (HF_HUB_OFFLINE/TRANSFORMERS_OFFLINE), credential-free, server bound to 127.0.0.1, harness on the same pod over loopback HTTP; Lium pod 08a1ff4a-248a-4ff2-948e-ea558f99e808 (huid cosmic-hawk-4f, node calm-fox-65), 1x NVIDIA H100 80GB HBM3, United States, USD 1.30/h
- Published source
- https://huggingface.co/Cloudflare/clef
Model and serving disclosure
Qwen/Qwen3.8-27B
Values come from the public v1.5.5 aggregate. Scores and ranks may change in a later release.
Read the full leaderboard, the v1.5.5 release page, and the published method.