Benchmarks
More
OverviewBenchmaxxingCompareChartsCost vs CapabilityProviders per ModelProvider explorerGatewaysEU & SovereignJev-class modelsImage JevBenchAbout
JevBench by Benchmark Heaven · v1.4.2.2 · individual system

LitJev (Qwen3.8-27B)

Jev rebuild · by Zhengxu Yu

JevBench v1.4.2.2 score

19.5

Rank #59 of 91 ranked systems.

43.8 points behind Jev 1.13.0's 63.3.

Where it sits among the 91 ranked systems. The marked tick is Jev 1.13.0 (63.3).

Published axes

intelligence
46.3
calibration
76.6
speed
66.7
cost
33.6

Against Jev 1.13.0

Four radars compare this fixed pair across the score axes, accuracy per tier, and accuracy by family on the hard tier and sealed set. Further out is better on every spoke.

  • A: LitJev — Jev rebuild · Score 19.5 (#59)
  • B: Jev 1.13.0 — Jev (TypeSafe, closed) · Score 63.3 (#4)

The four score axes

Radar: the four score axes, two systemsThe four score axes, LitJev vs Jev 1.13.0. Intelligence: 46.3 vs 53.1; Calibration: 76.6 vs 76.3; Speed: 66.7 vs 83.3; Cost: 33.6 vs 52.0.50100Intelligence46.3 · 53.1Calibration76.6 · 76.3Speed66.7 · 83.3Cost33.6 · 52.0
0–100, the values in the table. A label-only system has no calibration (counted as 0).

Accuracy per tier, incl. sealed

Radar: accuracy per tier, incl. sealed, two systemsAccuracy per tier, incl. sealed, LitJev vs Jev 1.13.0. Easy: 100% vs 100%; Standard: 98% vs 99%; Judge: 88% vs 95%; Hard: 73% vs 74%; Sealed: 31% vs 37%.50100Easy100% · 100%Standard98% · 99%Judge88% · 95%Hard73% · 74%Sealed31% · 37%
Share correct per tier; Sealed = the 308 private decisions, aggregate only.

Current question set by family (hard + sealed)

Radar: current question set by family (hard + sealed), two systemsCurrent question set by family (hard + sealed), LitJev vs Jev 1.13.0. Ambiguous / abstain: 53% vs 43%; Judge: 50% vs 54%; Long policy: 46% vs 44%; Multi-hop: 59% vs 64%; Probability: 40% vs 63%; Temporal / numeric: 24% vs 28%; Trade-off: 42% vs 55%; Routing: 100% vs 100%; Trap / adversarial: 88% vs 83%; Paraphrase: 29% vs 64%; Safety judge: 50% vs 38%.50100Ambiguous /abstain53% · 43%Judge50% · 54%Long policy46% · 44%Multi-hop59% · 64%Probability40% · 63%Temporal /numeric24% · 28%Trade-off42% · 55%Routing100% · 100%Trap /adversarial88% · 83%Paraphrase29% · 64%Safety judge50% · 38%
Share correct per family across the 220 hard-tier decisions (public and held out) and the 308 sealed decisions of v1.4, pooled; Routing is hard-tier only, Paraphrase and Safety judge sealed only.

Sealed set by family

Radar: sealed set by family, two systemsSealed set by family, LitJev vs Jev 1.13.0. Ambiguous / abstain: 41% vs 30%; Judge: 32% vs 34%; Long policy: 25% vs 28%; Multi-hop: 37% vs 45%; Paraphrase: 29% vs 64%; Probability: 21% vs 50%; Safety judge: 50% vs 38%; Temporal / numeric: 21% vs 29%; Trade-off: 23% vs 38%; Trap / adversarial: 58% vs 42%.50100Ambiguous /abstain41% · 30%Judge32% · 34%Long policy25% · 28%Multi-hop37% · 45%Paraphrase29% · 64%Probability21% · 50%Safety judge50% · 38%Temporal /numeric21% · 29%Trade-off23% · 38%Trap /adversarial58% · 42%
Share correct within each sealed family — system-level aggregates; the items stay private.

Availability and evidence

Openness
Code and weights marked open in the published row
License note
Apache-2.0 (code); Apache-2.0 base weights
Cost evidence
estimated; the board’s row disclosure contains the published basis.
Endpoint condition
our RunPod GPU (H100 NVL 96 GB, Canada), reached over the internet from Germany
Note on this row
The author's reproduction of Jev's decision layer on an off-the-shelf model, in its default configuration: Qwen3.8-27B, scores read from the output head, no training and no calibration file (its README says probabilities are not calibrated by default). Run serially on our GPU through an SSH tunnel, because its server binds to localhost; the request still crosses the internet and gets the ×2 + 0.15 s adjustment.
Published source
https://github.com/zhengxuyu/litjev

From the public v1.4.2.2 aggregate. Scores and ranks can change when a new release is published.

Read the full board and published method. The overall score is a composite, not raw accuracy.