Rnj-1 – 8B-parameter open-weight LLM optimized for code and STEM

Summary: Rnj-1 is an 8B-parameter open-weight dense model family offering base and instruction-tuned versions with strong code, tool-calling, math, and science capabilities comparable to leading open models. It supports local deployment and agentic workflows.

What it does

Rnj-1 provides base and instruction-tuned checkpoints optimized for coding, STEM tasks, and tool-calling, enabling agent behaviors and offline workflows.

Who it's for

Developers building agents or offline workflows requiring practical, small but capable models for code and STEM applications.

Why it matters

It delivers accessible, open-weight models optimized for specialized coding and STEM tasks, supporting local use and integration with agents.