Benchmarks
More
OverviewBenchmaxxingCompareChartsCost vs CapabilityProviders per ModelProvider explorerGatewaysEU & SovereignJev-class modelsAbout
JevBench by Benchmark Heaven · v1.4.2 · individual system

JevAct (einptein, jev1-2b-v2)

This detail uses the public, hash-checked v1.4.2 aggregate. Scores and ranks can change when a new release is published; the page preview remains name-only.

JevBench v1.4.2 score

16.9

Rank #65 of 89 ranked systems.

47.2 points behind Jev 1.13.0's 64.1.

Where it sits among the 89 ranked systems. The marked tick is Jev 1.13.0 (63.3).

Published axes

intelligence
29.3
calibration
55.1
speed
76.5
cost
64.3

Accuracy per tier, incl. sealed

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: JevAct — Jev rebuild · Score 16.9 (#65)
  • B: decider-4b v2 — system-one-open · Score 64.1 (#1)

The four score axes

Radar: the four score axes, two systemsThe four score axes, JevAct vs decider-4b v2. Intelligence: 29.3 vs 49.4; Calibration: 55.1 vs 75.0; Speed: 76.5 vs 92.9; Cost: 64.3 vs 60.9.50100Intelligence29.3 · 49.4Calibration55.1 · 75.0Speed76.5 · 92.9Cost64.3 · 60.9
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, JevAct vs decider-4b v2. Easy: 99% vs 100%; Standard: 69% vs 97%; Judge: 59% vs 88%; Hard: 37% vs 67%; Sealed: 23% vs 35%.50100Easy99% · 100%Standard69% · 97%Judge59% · 88%Hard37% · 67%Sealed23% · 35%
Share correct per tier; Sealed = the 308 private decisions, aggregate only.

Hard tier by family (v1.2 topics)

decider-4b v2 has no published v1.2 hard-tier family breakdown.

Radar: hard tier by family (v1.2 topics), two systemsHard tier by family (v1.2 topics), JevAct vs decider-4b v2. Adversarial: 75% vs —; Ambiguous: 43% vs —; Judge: 61% vs —; Long policy: 0% vs —; Multi-hop: 3% vs —; Probability: 40% vs —; Routing: 90% vs —; Temporal / numeric: 27% vs —; Trade-off: 58% vs —; Trap: 88% vs —.50100Adversarial75%Ambiguous43%Judge61%Long policy0%Multi-hop3%Probability40%Routing90%Temporal /numeric27%Trade-off58%Trap88%
Share correct within each family of the 47 v1.2 hard-tier decisions (public and held-out).

Sealed set by family

Radar: sealed set by family, two systemsSealed set by family, JevAct vs decider-4b v2. Ambiguous / abstain: 27% vs 43%; Judge: 34% vs 32%; Long policy: 0% vs 30%; Multi-hop: 21% vs 34%; Paraphrase: 14% vs 50%; Probability: 21% vs 36%; Safety judge: 31% vs 38%; Temporal / numeric: 23% vs 30%; Trade-off: 23% vs 35%; Trap / adversarial: 67% vs 33%.50100Ambiguous /abstain27% · 43%Judge34% · 32%Long policy0% · 30%Multi-hop21% · 34%Paraphrase14% · 50%Probability21% · 36%Safety judge31% · 38%Temporal /numeric23% · 30%Trade-off23% · 35%Trap /adversarial67% · 33%
Share correct within each sealed family — system-level aggregates; the items stay private.

Availability and evidence

Openness
Marked closed in the published row
License note
not stated (the author published only an endpoint and the evaluation client)
Cost evidence
estimated; the board’s row disclosure contains the published basis.
Endpoint condition
author-hosted endpoint (one machine at 115.190.124.200:18084, China), reached from a Hetzner server in Germany
Published source
https://github.com/fstandhartinger/jevbench/issues/66

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