Back to explore

Jevの実意思決定ベンチマーク公開:900件をOpus 5が採点

このベンチマークは900件の実際のJev意思決定をOpus 5が採点したもので、一致率は85.8/100。不一致はほぼJevが低信頼度を報告した箇所に集中し、誰でも自分のモデルで再実行できる。

Open weight builders will need a public yardstick. One went out today: 900 real Jev decisions graded by Opus 5, 85.8 of 100 agree, and the misses sit almost entirely where Jev reported low confidence. Anyone can rerun it against their own model:

· 0 likesOpen on X