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

Autoloops – Gemma 4 31B IT

Jev (TypeSafe, closed) · by Autoloops · API endpoint saw sealed item text

JevBench v1.4.2.2 score

30.0

Rank #44 of 91 ranked systems.

33.3 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
59.8
calibration
79.2
speed
84.0
cost
36.0

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.

The four score axes

Radar: the four score axes, two systemsThe four score axes, Autoloops – Gemma 4 31B IT vs Jev 1.13.0. Intelligence: 59.8 vs 53.1; Calibration: 79.2 vs 76.3; Speed: 84.0 vs 83.3; Cost: 36.0 vs 52.0.50100Intelligence59.8 · 53.1Calibration79.2 · 76.3Speed84.0 · 83.3Cost36.0 · 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, Autoloops – Gemma 4 31B IT vs Jev 1.13.0. Easy: 100% vs 100%; Standard: 99% vs 99%; Judge: 96% vs 95%; Hard: 83% vs 74%; Sealed: 46% vs 37%.50100Easy100% · 100%Standard99% · 99%Judge96% · 95%Hard83% · 74%Sealed46% · 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), Autoloops – Gemma 4 31B IT vs Jev 1.13.0. Ambiguous / abstain: 55% vs 43%; Judge: 82% vs 54%; Long policy: 58% vs 44%; Multi-hop: 68% vs 64%; Probability: 56% vs 63%; Temporal / numeric: 35% vs 28%; Trade-off: 61% vs 55%; Routing: 100% vs 100%; Trap / adversarial: 85% vs 83%; Paraphrase: 57% vs 64%; Safety judge: 44% vs 38%.50100Ambiguous /abstain55% · 43%Judge82% · 54%Long policy58% · 44%Multi-hop68% · 64%Probability56% · 63%Temporal /numeric35% · 28%Trade-off61% · 55%Routing100% · 100%Trap /adversarial85% · 83%Paraphrase57% · 64%Safety judge44% · 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, Autoloops – Gemma 4 31B IT vs Jev 1.13.0. Ambiguous / abstain: 46% vs 30%; Judge: 73% vs 34%; Long policy: 33% vs 28%; Multi-hop: 53% vs 45%; Paraphrase: 57% vs 64%; Probability: 46% vs 50%; Safety judge: 44% vs 38%; Temporal / numeric: 29% vs 29%; Trade-off: 42% vs 38%; Trap / adversarial: 50% vs 42%.50100Ambiguous /abstain46% · 30%Judge73% · 34%Long policy33% · 28%Multi-hop53% · 45%Paraphrase57% · 64%Probability46% · 50%Safety judge44% · 38%Temporal /numeric29% · 29%Trade-off42% · 38%Trap /adversarial50% · 42%
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
Gemma terms of use (Google) for the base weights; hosted Autoloops service
Cost evidence
measured; the board’s row disclosure contains the published basis.
Endpoint condition
Autoloops production API, reached from a Hetzner server in Germany
Note on this row
Gemma 4 31B IT served through the Autoloops systemone API. Cost = the API's own token usage x Autoloops' published rates ($0.20/M input, $0.45/M output; no output tokens were used). API measurement: public and sealed item content reached the operator endpoint; no gold labels or answers were sent.

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.