OpenRouter clocks Jev over 5x faster than next model
Jev's slowest requests still beat every other model's median, OpenRouter says.
OpenRouter tested Jev, the decision model from TypeSafe AI, against popular LLMs on its platform using Ori Eval, an evaluation tool for judging tasks. The result, posted by OpenRouter, Jev was more than 5x faster than the next fastest model, and even its slowest requests beat every other model's median.
Jev was >5x faster than the next fastest model, and even its slowest requests beat every other model's median.
That is the whole result as published. OpenRouter did not share accuracy numbers, pricing, or which models it tested against, so there is no basis for comparing cost or correctness here, only speed.
OpenRouter has separately described Jev as answering yes or no and multiple choice questions with a confidence score, and called it 10x cheaper and faster than an LLM in its own explainer of what a decision model is. For people building on it, this is one vendor's speed measurement on judging, not a full benchmark, so the accuracy and cost tradeoffs remain unmeasured until someone publishes them.