Nace AI Releases Drex 1.5 for Option Scoring in Single Pass

Nace AI has made Drex 1.5 available as open source, introducing a specialized nine-billion-parameter model designed for decision-making tasks. According to MarkTechPost, this release marks a shift toward models that prioritize scoring specific options over traditional text generation. The model is built to return a probability for every available option in a single forward pass, streamlining the evaluation process for developers who need rapid, deterministic outputs.

How does Drex 1.5 handle decision-making differently?

Unlike standard language models that generate sequential text, Drex 1.5 focuses on evaluating predefined choices. MarkTechPost reports that the architecture allows the model to process input and output a probability score for each option simultaneously. This single-pass approach eliminates the need for iterative generation steps, potentially reducing latency and computational overhead for applications that require quick comparative analysis of multiple distinct outcomes.

View the original post on X

What are the technical specifications of the release?

The Drex 1.5 model contains nine billion parameters and supports a context window of up to 128K tokens. MarkTechPost notes that the model was evaluated using Decision Index 0.3.1, where it achieved a score of 58.08. These specifications indicate that the model is capable of handling large volumes of contextual information while maintaining the structural integrity required for precise option scoring, making it suitable for complex decision-support systems that require extensive data intake.

Who is the developer behind this model?

Nace AI is the organization responsible for developing and open-sourcing Drex 1.5. By releasing the model publicly, the company allows developers and researchers to access, modify, and deploy the architecture for various decision-scoring applications. This open-source strategy aligns with broader industry trends toward democratizing access to specialized AI models, enabling a wider community to build upon the foundational work provided by Nace AI.

What comes next for developers using this model?

Developers can now integrate Drex 1.5 into their workflows to evaluate options based on probability scores. Since the model is open source, teams can audit the architecture and fine-tune it for specific use cases. The availability of the model on public repositories allows for immediate experimentation, enabling engineers to test its performance against their existing decision-making pipelines without waiting for proprietary API access.

What to do and how to stay safe: Nace AI

• Review the open-source repository for Drex 1.5 to understand the model architecture and licensing terms before integration.
• Test the model’s performance on your specific dataset to verify that the 58.08 Decision Index score translates to your use case.
• Ensure your infrastructure can handle the nine-billion-parameter size and the 128K token context window requirements.
• Monitor community updates and issues in the repository for any security patches or performance improvements released by Nace AI.

Step-by-step guide: How AI Risk Management Works: The Hidden Failure Points

General security guidance from the Sandbox Dispatch newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

What is the parameter count of Drex 1.5?

Drex 1.5 is a nine-billion-parameter model, according to MarkTechPost.

How does Drex 1.5 output results?

It returns a probability for every option in a single forward pass, rather than generating text.

What score did Drex 1.5 achieve on Decision Index 0.3.1?

The model scored 58.08 on Decision Index 0.3.1, as reported by MarkTechPost.