Danielbrdz/Barcenas-31b-Fable

VISIONConcurrency Cost:2Model Size:31BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:Jun 22, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

Danielbrdz/Barcenas-31b-Fable is a 31 billion parameter causal language model based on Google's Gemma 4 31b architecture, featuring a 32768 token context length. It is specifically fine-tuned with high-quality data from Claude Fable 5 for agentic use cases. This model is designed for efficient local and private deployment on a single GPU, making it suitable for agent-based applications.

Loading preview...

Barcenas-31b-Fable: An Agentic LLM

Barcenas-31b-Fable is a 31 billion parameter language model developed by Danielbrdz, built upon Google's Gemma 4 31b architecture. This model represents Danielbrdz's largest and most powerful creation to date, distinguished by its specialized training for agentic applications.

Key Capabilities & Training

  • Base Model: Utilizes the robust Gemma 4 31b instruction-tuned foundation from Google.
  • Specialized Fine-tuning: Trained extensively with the agentic-distill-fable-5-sft dataset, which incorporates high-quality data directly from Claude Fable 5.
  • Agentic Use: Optimized specifically for tasks requiring agent-like behavior and reasoning.
  • Local Deployment: Designed to run efficiently on a single GPU, enabling local and private execution of agentic workloads.

Use Cases

This model is particularly well-suited for:

  • Developing and deploying AI agents locally.
  • Applications requiring private and efficient agentic processing.
  • Tasks that benefit from a powerful LLM capable of running on consumer-grade hardware (with sufficient VRAM).

Its focus on agentic use and local deployability differentiates it from more general-purpose models, offering a specialized solution for developers building agent-based systems.