Jev beats Gemini 2.5 Flash Lite on classifier eval quality and speed
Malte Ubl ran TypeSafe AI's Jev against an existing classifier eval, where it saturated the eval on quality and was 6x faster than Gemini 2.5 Flash Lite.
Jev classified 300 web pages into 7 page types, with 88 matching Opus 5 results, at 61 ms per page, and a side-by-side replay of both models is provided.
Classifying node types is the same shape as what I ran: Jev sorted 300 web pages into 7 page types and 88 of 100 matched Opus 5. At 61 ms a page the animation budget stays yours. Side by side replay of both models: