Nace AI Open-Sources Drex 1.5: A 9B Decision Model That Scores Options, Not Text
A 9B open-weight model that scores options, not text—Drex 1.5 trades generative output for decision speed and measurability.
Why it matters
Nace AI's Drex 1.5 represents a specific capability trade: it returns probability scores for a fixed set of options in a single forward pass, rather than generating free-form text. This is useful for structured decision workflows where the output space is known and measurable (e.g., routing, ranking, classification at scale). The 58.08 score on Decision Index 0.3.1 and 128K context window are the key claims; practitioners deploying agents or agentic workflows that need fast, deterministic option scoring should note the capability and benchmark.
The key facts
8 to knowModel size: 9B parameters, open-sourced
Architecture: returns probability score for every option in 1 forward pass (not generative text)
Benchmark: 58.08 on Decision Index 0.3.1
Context window: 128K tokens
Source: MarkTechPost (reporting vendor announcement, not independent testing)
Model: Drex 1.5, 9B parameters, open-sourced
Architecture: scores probability for each option in one forward pass (no iterative text generation)
Source: MarkTechPost (secondary; no primary release link provided in article)
The story so far
Earlier coverage of this storyline
Go to the source
MarkTechPostmarktechpost.com
Publisher excerpt: Nace.AI has open-sourced Drex 1.5, a 9B decision model that returns a probability for every option in 1 forward pass. It scores 58.08 on Decision Index 0.3.1 and reads up to 128K tokens.