Open-Jev 27B v1.1 (Zefan Cai, LoRA + decision head on Qwen3.8-27B)
Data: JevBench v1.6.1, published 2026-10-06
Capability 68.0 · Not ranked in the current release
cost 22.2× Jev (v1.5 reference); latency 2.8× Jev (v1.5 reference)
Last measured: 2026-10-04 · measurement release: v1.6.1
Apache-2.0
intelligence
52.5
calibration
83.5
speed
72.5
cost
14.4
Composite (secondary)
2.9 · Composite rank #94
Cost and measurement conditions
$0.716 per 1,000 decisions (estimated)
latest published v1.5 row (v1.5.8 combined file) cost per 1,000 decisions carried (pricing rules unchanged; v1.6 item lengths differ): Frozen BASE_REFERENCES Qwen/Qwen3.8-27B market price (USD 0.42/M input, 3.00/M output) applied to 2,767,510 input tokens and zero output tokens over 1,624 decisions; self-hosted estimate, not measured hosted billing. Lineage verified by architecture; no pricing gate. The author method encodes each candidate independently.
Measured p50 latency: 0.790 s (on-GPU, no network, loopback; not directly comparable with API latency, which we time end to end from our measuring host, network included)
Ranking and cap use: 1.73 s (est. incl. network (2x + 150 ms))
Self-hosted model: measured p50 0.790 s on-GPU, no network (timed on our GPU over loopback). Not directly comparable with API latency, which we time end to end from our measuring host, network included. The ranking and latency cap use the same estimate for every self-hosted row: 2 × measured + 150 ms = 1.73 s. Model authors can ask us to time a hosted endpoint instead (on request).
x2 + 0.15 s (assumption, not measured)
Self-hosted author loader, Transformers reference path on RTX PRO 6000 Blackwell; speed is a LOWER BOUND: flash-linear-attention and causal-conv1d kernels were not installed. The author documents the kernel path as answer-identical. Each candidate is encoded independently, producing high input-token usage.