1. Start with the official public sources
If you need concrete public model information, begin with the DeepBrainz Hugging Face organization and then use DeepBrainz Labs for research guidance. The main corporate site explains how the model line fits the product system, but Hugging Face and Labs are the deeper sources for model-facing questions.
2. Supported and experimental are not the same thing
The R-series material explicitly separates supported releases, long-window variants, research checkpoints, and community quantizations. That distinction matters because support expectations should match the artifact type.
- Supported releases are the safest starting point for serious evaluation.
- Experimental variants are useful for exploration, not automatic production assumptions.
- Research checkpoints help explain the direction, but they are not the same thing as recommended starting points.
- Community builds can be useful, but support and reliability expectations differ.
3. Current public R1 line
The current public DeepBrainz R1 line highlights three recommended starting points: R1-4B, R1-2B, and R1-0.6B-v2. Public materials frame them as agentic models for frontier agent systems in production, with long-window and research variants clearly separated.
4. How R1 fits the product stack
DeepBrainz-R is the model and agent-systems layer. Lexopedia AI is the current production workspace for research and technical work. AgentFoundry is the reviewed execution layer. Labs explains the research logic and evaluation side behind that stack.
5. What support can help with
Support can help you find the right official source, clarify supported versus exploratory public artifacts, and route you to the right DeepBrainz surface. Support is less useful if the question assumes private guarantees that are not present in the public release notes or official docs.