Drex v2.0
Drex v2.0 is Nace.AI's larger decision-only AI following Drex v1.5: it reads a situation written as text or JSON and returns probabilities for a set list of choices, a yes/no answer, or a step on a rating scale. It is built on an Alibaba Qwen model; the whole model is large but uses only a small part at a time, so it runs lightly for its size. Nace.AI reports a higher score than Drex v1.5 on a public decision test and more wins in games played against Jev. In our tests it beat Drex v1.5 on Korean and Japanese judgment accuracy and in Tetris, but in time-pressured Pong its late answers left it far behind. You can download the public weights and run it yourself, given a lot of storage; it comes under the same Nace.AI license as Drex v1.5, which counts evaluation as use and requires larger companies, or those offering competing products, to get separate permission. Nace.AI says it scores below Jev on knowledge and reasoning questions.
API cost
Expert Performance Evaluation
Classification and judgment
All classification and judgment AI rankingsSee each evaluation’s detail page for its source and last updateData sources & removal requests