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

jqv (Qwen3-32B zero-shot)

Jev rebuild · by hjmurmur (Octalab)

JevBench v1.4.2.2 score

44.4

Rank #19 of 91 ranked systems.

18.9 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
46.4
calibration
71.6
speed
74.6
cost
47.5

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: jqv — Jev rebuild · Score 44.4 (#19)
  • 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, jqv vs Jev 1.13.0. Intelligence: 46.4 vs 53.1; Calibration: 71.6 vs 76.3; Speed: 74.6 vs 83.3; Cost: 47.5 vs 52.0.50100Intelligence46.4 · 53.1Calibration71.6 · 76.3Speed74.6 · 83.3Cost47.5 · 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, jqv vs Jev 1.13.0. Easy: 100% vs 100%; Standard: 96% vs 99%; Judge: 92% vs 95%; Hard: 65% vs 74%; Sealed: 28% vs 37%.50100Easy100% · 100%Standard96% · 99%Judge92% · 95%Hard65% · 74%Sealed28% · 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), jqv vs Jev 1.13.0. Ambiguous / abstain: 39% vs 43%; Judge: 54% vs 54%; Long policy: 31% vs 44%; Multi-hop: 45% vs 64%; Probability: 35% vs 63%; Temporal / numeric: 26% vs 28%; Trade-off: 39% vs 55%; Routing: 90% vs 100%; Trap / adversarial: 88% vs 83%; Paraphrase: 36% vs 64%; Safety judge: 56% vs 38%.50100Ambiguous /abstain39% · 43%Judge54% · 54%Long policy31% · 44%Multi-hop45% · 64%Probability35% · 63%Temporal /numeric26% · 28%Trade-off39% · 55%Routing90% · 100%Trap /adversarial88% · 83%Paraphrase36% · 64%Safety judge56% · 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, jqv vs Jev 1.13.0. Ambiguous / abstain: 30% vs 30%; Judge: 34% vs 34%; Long policy: 13% vs 28%; Multi-hop: 24% vs 45%; Paraphrase: 36% vs 64%; Probability: 18% vs 50%; Safety judge: 56% vs 38%; Temporal / numeric: 27% vs 29%; Trade-off: 27% vs 38%; Trap / adversarial: 58% vs 42%.50100Ambiguous /abstain30% · 30%Judge34% · 34%Long policy13% · 28%Multi-hop24% · 45%Paraphrase36% · 64%Probability18% · 50%Safety judge56% · 38%Temporal /numeric27% · 29%Trade-off27% · 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 (Qwen3-32B weights); serving code public
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
A stock Qwen3-32B with no decision training: the state is prefilled once, each question is an isolated branch and the answer is read from the option-letter logits, with one fitted temperature (3.02, 400 MMLU validation items). Re-run in v1.2.8 on our own GPU from the now-public serving code (Octalab-Inc/jqv 0189b67), so all 534 decisions including the held-out hard items were asked; this full run replaces the v1.2.7 partial row, which had been measured on the submitter's machine. Cost is the base model's public per-token tariff, not free.
Published source
https://github.com/Octalab-Inc/jqv

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.