Key Info

OpenRouter tested Jev, a decision model by typesafeai, using Ori Eval against popular LLMs on judging tasks. Jev was more than 5x faster than the next fastest model, and even its slowest requests beat every other model's median speed.

Highlights

  • Jev is a decision model that answers yes/no and multiple-choice questions with a confidence score, designed for breaking software development into small decisions.
  • In OpenRouter's benchmark, Jev outperformed all tested LLMs on speed, with even its worst-case latency faster than the median of competing models.
  • The test used Ori Eval on OpenRouter, and supplementary mini-app examples with code are available in OpenRouter Labs.