ModelsBenchmark resultEvaluation & Benchmarks · Large Language Models1 source · Oct 10, 2026

Nace AI Open-Sources Drex 1.5: A 9B Decision Model That Scores Options, Not Text

Nace.AI has open-sourced Drex 1.5, a 9B decision model for agents and backend workflows.

Proof1 independent outlet

Key points

  • The Drex 1.5 decision model does not write text.
  • It reads a state and typed questions, then returns a probability for every option.
  • Nace reports 58.08 on the public Decision Index 0.3.1, the top score under 10B parameters.
  • Drex 1.5 is a decision model from Nace.AI that scores a fixed set of options in 1 forward pass.

Sources (1)

Extractive summary: sentences quoted from the sources.

Hype check

Independently evaluated4 independent looks since

  1. Evaluation · Oct 10, 2026The Company a Question Keeps: Auditing Open Jev-Style Models for Retrieval
  2. Evaluation · Oct 10, 2026SPARE: Shared Perception and Adaptive Retrieval of Evidence for Streaming Video Understanding
  3. Evaluation · Oct 10, 2026Verify, Act, Distill: Turning Test-Time Compute into Test-Time Adaptation for Vision-Language-Action Models
  4. Evaluation · Oct 10, 2026SEEK: Agentic Knowledge Graph Retrieval for Secure Code Generation
  5. Follow-up · Oct 10, 2026Fine-Tuning Strategy, Not Just Data: A Statistically Rigorous Audit of Gender and Stereotypical Bias in Embedding Models for HR Recommendation Systems
  6. Follow-up · Oct 10, 2026[AINews] TypeSafe/Jev at >$100M ARR, $7.5B valuation 3 weeks after launch

Before this

  1. Oct 9, 2026LiquidAI/LFM2.5-24B-A2B
  2. Oct 9, 2026Qwen/Qwen-Image-2.1-Turbo
  3. Oct 9, 2026Introducing Clef-omni with full multimodality, plus a faster Clef and a cheaper Clef-flash
  4. Oct 8, 2026Show HN: Jevman – AI decision models play Pac-Man
  5. Oct 8, 2026nvidia/Qwen3.8-Flash-Next-FP4-Local
  6. Oct 8, 2026unslothai/unsloth v0.1.905-beta: Sandboxing is here!

Related