yev2 Author Benchmark - JevBench hard111: leaderboard
yev2 Author Benchmark - JevBench hard111 — Author calibrated letter-logit readout via yev serve; native BF16 and separately released FP8 artifact, same benchmark table. Existing DecideBench v2 headline covered. Source does not characterize JevBench as held-out; yev1 baseline is copied from preceding release. Metric: Accuracy (%). Source: https://huggingface.co/choyiny/yev2-35b-a3b
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Rough model projection: around 2026. 2 models tracked. Updated 16 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | yev2-35B-a3B | 81.1 | |
| 2 | yev2-35B-a3B [FP8] | 78.4 |
Interactive version: theaggregate.ai/benchmark?slug=yev2-author-benchmark-jevbench-hard111 · How It Works · Data refreshed daily, snapshot 2026-11-16.