JevBench Debuts: Nine Models Tested Across 242 Decisions, Top Accuracy 97.1%
JevBench v1, a public benchmark for typed-decision "Jev-class" models, evaluated nine models across 242 decisions with top accuracy of 97.1%. Separately, a developer is building the first Jev benchmark to compare official Jev against open-source replicas under the benchmarkheaven project.
2026-09-19 ~ 2026-09-19 · 2 related posts
- First benchmark to compare official Jev against open-source lookalikes in the works — airesearch12 · 2026-09-19
- JevBench v1 puts nine typed-decision models head-to-head across 242 decisions — airesearch12 · 2026-09-19