Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Nace AI details Drex training and updates Drex Max v1.1

AI summary

Nace AI published a technical report covering the architecture, training, evaluation, calibration and local inference of four open-weight Drex decision models. The models use full-parameter supervised fine-tuning on 1.15 million verified decision records, followed by consistency training and reinforcement learning from adjusted feedback; Nace AI also released Drex Max v1.1. Nace AI claimed that Max v1.1 reached the state of the art and Pareto frontier for decision models, while Artificial Analysis reportedly ranked it first, but these performance claims remain vendor-attributed and the v1.1 weights were described as still forthcoming.

In-depth insight

The report provides unusually detailed training and deployment guidance for models designed for agent routing and tool selection, while the Max update is a material addition to the open-weight Drex lineup.

#Nace AI#Drex Max v1.1#Drex Max#Drex DLM#Drex v1.5#Drex v2.0#Artificial Analysis#OpenRouter#Ollama#llama.cpp#Kapitsa

Original · 2 min readRead the original
Group chat

No comments yet. Start the conversation.