Loading original post…Open on X ↗
Jev-1.13 classified 3,373 à-la-carte dishes into food types in 9.7s for $0.11 (2.56M input tokens), hitting 84.5% agreement with strictly typed labels—misses came from polysemous dish names, never hallucinated categories.
Original post
Loading original post…Open on X ↗
jev
Start typing to find a use case.