Drex: how Nace.AI built a fast, calibrated decision model

Decision models return calibrated probabilities over your options in one forward pass, avoiding output parsing and long generation. Drex Max answers in about 183 ms versus seconds for frontier LLMs, and only 4.3% of answers flip when options are reordered.

• open weights — A model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.
• calibration — How well a model's confidence matches reality — a calibrated model saying "90% sure" is right about 90% of the time.

• articleNace AI ships Drex Max, a multimodal decision model for computer-use agents
• articleNace AI open-sources Drex 2.0 decision model weights
• articleNace AI open-sources Drex, a diffusion-LM decision model

Checking sign-in…

Loading comments…