JevBench v1.5: Cygnet and Winnow-12B virtually tied for the top with no solo #1
airesearch12 · x · 2026-09-29
Benchmark Heaven's JevBench v1.5 is out, and for the first time there's no clear solo #1: Cygnet and Winnow-12B are so close they essentially share the top spot, while the original "chad Jev" falls to third. The update brings many methodological improvements, scoring 1,624 decisions per system (904 open, 720 sealed) with half from Intelligence tasks, and now scores Choice, Noul and Score equally.
JevBench is a meme-flavored "Jev-class model" leaderboard; the site also offers cost-vs-capability comparisons and provider region filtering.
Related event: JevBench v1.5 Released: Cygnet and Winnow-12B Tie at the Top(2 posts)→
More from Fun
- Villain-boss POV video made with Magnific and Seedance 2.5, prompt shared — charis_ai · 2026-09-29
- Earth Rover Challenge: robots run policies in the wild in Costa Rica, dodging curious kids — micoolcho · 2026-09-29
- Claude loves 'never' over 'not': 693 occurrences in one monorepo — tulensrma · 2026-09-29
- Documentary featuring photographer zemotion nominated for an International Emmy — zemotion · 2026-09-29
- The 'original API' meme: rate-limited, needs coffee, still undefeated at touching grass — SucceededMind · 2026-09-29
- Asked about copyright, Gemini researched how to get an annoying person to stop talking — Nice_Assumption_6396 · 2026-09-29