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

Decision 2B (FlyMy.AI, v59)

Jev rebuild · by FlyMy.AI (@denti)

JevBench v1.4.2.2 score

35.8

Rank #31 of 91 ranked systems.

27.5 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
38.8
calibration
74.1
speed
84.3
cost
62.5

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: Decision 2B — Jev rebuild · Score 35.8 (#31)
  • 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, Decision 2B vs Jev 1.13.0. Intelligence: 38.8 vs 53.1; Calibration: 74.1 vs 76.3; Speed: 84.3 vs 83.3; Cost: 62.5 vs 52.0.50100Intelligence38.8 · 53.1Calibration74.1 · 76.3Speed84.3 · 83.3Cost62.5 · 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, Decision 2B vs Jev 1.13.0. Easy: 100% vs 100%; Standard: 88% vs 99%; Judge: 70% vs 95%; Hard: 58% vs 74%; Sealed: 26% vs 37%.50100Easy100% · 100%Standard88% · 99%Judge70% · 95%Hard58% · 74%Sealed26% · 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), Decision 2B vs Jev 1.13.0. Ambiguous / abstain: 37% vs 43%; Judge: 46% vs 54%; Long policy: 31% vs 44%; Multi-hop: 38% vs 64%; Probability: 44% vs 63%; Temporal / numeric: 21% vs 28%; Trade-off: 37% vs 55%; Routing: 80% vs 100%; Trap / adversarial: 73% vs 83%; Paraphrase: 43% vs 64%; Safety judge: 44% vs 38%.50100Ambiguous /abstain37% · 43%Judge46% · 54%Long policy31% · 44%Multi-hop38% · 64%Probability44% · 63%Temporal /numeric21% · 28%Trade-off37% · 55%Routing80% · 100%Trap /adversarial73% · 83%Paraphrase43% · 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, Decision 2B vs Jev 1.13.0. Ambiguous / abstain: 24% vs 30%; Judge: 29% vs 34%; Long policy: 18% vs 28%; Multi-hop: 16% vs 45%; Paraphrase: 43% vs 64%; Probability: 36% vs 50%; Safety judge: 44% vs 38%; Temporal / numeric: 25% vs 29%; Trade-off: 23% vs 38%; Trap / adversarial: 25% vs 42%.50100Ambiguous /abstain24% · 30%Judge29% · 34%Long policy18% · 28%Multi-hop16% · 45%Paraphrase43% · 64%Probability36% · 50%Safety judge44% · 38%Temporal /numeric25% · 29%Trade-off23% · 38%Trap /adversarial25% · 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 notices on the included code and the pinned base; the weights are an evaluation preview under EVALUATION-PERMISSION.md, not a cleared commercial release
Cost evidence
estimated; the board’s row disclosure contains the published basis.
Endpoint condition
our RunPod GPU (L40 48 GB, Czechia), reached over the internet from Germany
Note on this row
Flymy-ai/decision-2b-preview revision df57b75db927acc9ad91ec8115508c1e487086eb (checkpoint minicpm5_reduced_v16_4k_v59), base openbmb/MiniCPM5-2B revision 12a3808a956f869c767195e9266b59c4d21d92e2, the submitter's own FlyMyJevPackageAdapter and frozen calibrator, bf16, unmerged adapter, 4096-token packing, transformers 4.57.6 / peft 0.15.2 as pinned, torch 2.8.0 from the pod image, on our RunPod L40 in Czechia Offline local open-weight inference on the frozen 308-item v1.4 set in a network-disabled, read-only container; no operator endpoint received sealed text.
Published source
https://huggingface.co/flymy-ai/decision-2b-preview

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.