Developers and researchers wanting a small open model for coding agents; free weights, though the 32k context is modest.
Kenning's take
What it does
Rnj-1 is a pair of open base and instruction-tuned language models from Essential AI, with 8 billion parameters and a 32,000-token context window, loosely following Google's Gemma 3 design. Essential AI says it compares well with open models of similar size on coding, agentic software tasks such as SWE-bench, tool use and maths. The weights are released openly, and the Instruct model is also offered as a hosted API on Together AI.
- 8B base and instruct models
- 32k context window
- Agentic coding and SWE-bench focus
- Function calling and tool use
- Holds quality when quantised to FP8 and NVFP4
Who it suits. Developers and researchers who want a small open model for coding agents and reasoning tasks.
Key facts
- Category
- AI models
- Made by
- Essential AI
- Based in
- United States
- Released
- 5 December 2025
- How to get it
- open weights (base and instruct model cards), and as Rnj-1 Instruct on Together AI's API
- Works on
- Open weights, API (Together AI)
- In the UK
- Yes
- Free plan
- open weights free to download
FAQ
When was Rnj-1 released?
Essential AI announced Rnj-1 on 5 December 2025.
Is Rnj-1 open source?
Essential AI describes it as its first contribution to open-source AI, releasing both base and instruct models.
How big is Rnj-1?
It has 8 billion parameters and a 32,000-token context window.
Facts from the maker's own pages and the sources listed. Checked 2026-10-11.
Similar tools
ACE-Step 1.5
Open-source music generation model that writes full songs with lyrics and runs on ordinary consumer GPUs.
AlphaGenome Atlas
Google DeepMind database of AI-predicted effects for every possible single-letter change in human DNA.
Amazon Nova 2
Amazon's second family of Nova AI models, covering reasoning, speech conversation and combined text-and-image generation.
Cisco Antares
Cisco's small open-weight models that find where known security vulnerabilities sit in a codebase, running locally.