JevBench

As of Oct 14, 2026, JevBench (v1.6.1) is maintained by Benchmark Heaven. Metric capability or headline composite. Ceiling 100. Saturation unknown. Contamination risk med.

Maintainer
Benchmark Heaven
Metric
capability or headline composite
Ceiling
100
Saturation
unknown
Contamination
med
Cadence
Declared cadence: revisioned. Catalog snapshots: 1, latest Oct 8, 2026.
Source
JevBench
Categories

Where it counts

Top 5

Current top scores

  1. Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) · 82.79186839726479 · Oct 8, 2026
  2. Quyet-1.0-Large (Chinh Nguyen, Gemma-4-31B decoder, option-letter logits) · 81.68880454670384 · Oct 8, 2026
  3. torchcast-decision-27b (Torchcast AI, Qwen3.8-27B fine-tune) · 81.56565836114741 · Oct 8, 2026
  4. decisio v0.8.0 on gemma-4-31B-it (frozen, FP8 on load, one prefill per state, self-hosted) · 79.56974174390604 · Oct 8, 2026
  5. Kev 27B (Kev 1.0, Qwen3.8-27B) · 78.93896122956639 · Oct 8, 2026
ModelScoreDateKind
Sage 1.3.0 (Levanto Labs)74.049921791754682026-10-08snapshot
H2O-Lightning-4B v1.1 (H2O.ai, Qwen3.5-4B fine-tune, stock vLLM + open shim)74.996196369350472026-10-08snapshot
Mercury Decide (Inception; System One decisions API, served free on OpenRouter as inception/mercury-decide:free)72.368240496377282026-10-08snapshot
decisio v0.8.0 on gemma-4-31B-it (frozen, FP8 on load, one prefill per state, self-hosted)79.569741743906042026-10-08snapshot
Jev 1.13.0 (TypeSafe AI)71.491186392845142026-10-08snapshot
Quyet-1.0-Large (Chinh Nguyen, Gemma-4-31B decoder, option-letter logits)81.688804546703842026-10-08snapshot
decider-12b v2 (Mapika)72.500476272917332026-10-08snapshot
wity-1 (Wity, reasoning auto)70.765340103310632026-10-08snapshot
torchcast-decision-12b (Torchcast AI, Gemma-4-12B fine-tune, option-letter logprob readout)71.738946113239242026-10-08snapshot
Winnow-12B71.245861762500372026-10-08snapshot
deck-31B (krishna765, frozen Gemma-4-31B-it, TorchAO FP8 dynamic)77.59074291252612026-10-08snapshot
Cygnet (blockbrain, frozen Gemma-4-12B-it)70.903410555155652026-10-08snapshot
decisio v0.8.0 on gemma-4-12B-it (frozen, one forward pass, self-hosted)70.741466143691442026-10-08snapshot
decisio v0.9.0 on gemma-4-12B-it (frozen, one prefill per question, self-hosted)69.954380228328862026-10-08snapshot
Jev-Omni (akhilaaa3, Gemma-4-12B merged)71.28736627859182026-10-08snapshot
Xor 26B-A4B (Juspay, Gemma-4-26B-A4B, bf16)72.406519859769652026-10-08snapshot
Bobcat Flash 1.2 (Gemma-4-26B-A4B-it + two merged rank-64 LoRA adapters, typed-decision readout at the first answer position, FP8, self-hosted)73.49367227371442026-10-08snapshot
SPX-CD Flash (SurdAI, hosted /v1/systemone, Oct-4 checkpoint)66.680171553862122026-10-08snapshot
decider chat on Gemma-4-31B-it (Mapika, frozen base, inference technique)70.582676784055542026-10-08snapshot
Surogate Rune 26B-A4B v3 (v1.6 pool)73.468079721406572026-10-08snapshot
GEV-26B-Decide (AutoTrust, Gemma-4-26B-A4B + LoRA + head)70.076131102590722026-10-08snapshot
SPX-CD-Omni (SurdAI, google/gemma-4-12B-it + LoRA r32, one forward pass, self-hosted)68.12866240517342026-10-08snapshot
Diffusion Jev (DiffusionGemma 26B-A4B on patched SGLang, 48-step diffusion readout, self-hosted)58.087508127951112026-10-08snapshot
René-1 31B FP8 (salfatigroup, Gemma 4 31B full fine-tune, one-pass option readout)75.95408331830042026-10-08snapshot
Bespoke Nimble 9B v3 (Qwen3.5-9B LoRA)66.792908804909932026-10-08snapshot
APUS-OpenJev-v1-9B (merged bf16 checkpoint-3000 on Qwen3.5-9B, their own native runtime, full depth, self-hosted)56.9078450990855062026-10-08snapshot
Plumb-4B (crh225, JevK5 v0.2 + LoRA)65.421844228319232026-10-08snapshot
SPX-CD Pro (SurdAI, hosted /v1/systemone, Oct-4 checkpoint)47.669708207063672026-10-08snapshot
Messier One v0.2 (Qwen3.5-4B fine-tune, one prefill, self-hosted)58.1305624437446442026-10-08snapshot
spark-s1-4b-v6 (Open Spark Jev, abhishek085)54.0160432777781242026-10-08snapshot