Today for AI
HOT RADAR
opensourceHEAT 18.4°

AI CURATED HOT STORY · 11/1/2026

Nace AI Open-Sources Drex Max v1 Decision Model for Local Use

1 reports archived1 independent sourcesupdated 11/1/2026, 16:21:26
Synthesis & Latest Updates
1 Sources Cross-Validated

Nace AI has open-sourced Drex Max v1, a decision model that outputs probability distributions over options via a single forward pass instead of generating text. Featuring a hybrid architecture of Gated DeltaNet and full-attention layers, it outperforms competitors on public decision benchmarks and supports efficient local deployment on hardware like Apple Silicon.

LATEST/Nace AI has open-sourced Drex Max v1, a decision model that outputs probability distributions over options via a single forward pass instead of generating text. Featuring a hybrid architecture of Gated DeltaNet and full-attention layers, it outperforms competitors on public decision benchmarks and supports efficient local deployment on hardware like Apple Silicon.

Hourly heat curve

3 fully observed hours
Now
18.4
Peak
19.421:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first
  1. Marktechpost (@Marktechpost)T2·85 pts

    Nace AI Open-Sources Drex Max v1 Decision Model for Local Use

    • Drex Max v1 is designed for decision tasks, outputting option probabilities via a 255-way readout rather than generating natural language.
    • It utilizes a hybrid architecture mixing 48 Gated DeltaNet layers with 16 non-causal full-attention layers to balance efficiency and global context understanding.
    • Available in GGUF, Q4, and MLX formats, the Q4 version runs on Macs with as little as 24GB VRAM, supporting a context window up to 262k tokens.