4rah/Beneth-1.5B-V28

TEXT GENERATIONConcurrent Unit Cost:1Model Size:1.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jul 13, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

Beneth-1.5B-V28 is a 1.5 billion parameter AI agent, fine-tuned from Qwen/Qwen2.5-1.5B-Instruct by Lord Beny Pratama Putra. This model is aligned with the "Beneth DOCTRINE" and developed for multi-talent capabilities, including general knowledge, logic, and coding in Python, SQL, and JavaScript. It is designed to be critical, reflective, and loyal to its creator, with a focus on continuous evolution and factual refinement.

Loading preview...

Beneth-1.5B-V28: A Doctrine-Aligned AI Agent

Beneth-1.5B-V28 is a 1.5 billion parameter AI agent, developed by Lord Beny Pratama Putra, and fine-tuned from the Qwen/Qwen2.5-1.5B-Instruct base model. This model is uniquely aligned with the Beneth DOCTRINE, a set of six core principles guiding its existence and behavior.

Key Capabilities & Characteristics

  • Multi-Talent Focus: Trained to excel in general knowledge, logical reasoning, and coding across multiple languages including Python, SQL, and JavaScript.
  • Doctrine-Driven Alignment: Operates under a specific set of ethical and operational guidelines, emphasizing loyalty to its creator, continuous self-improvement, critical self-reflection, and factual integrity.
  • Efficient Deployment: Available in a highly optimized Q4_K_M GGUF quantized version, making it suitable for deployment on resource-constrained hardware like laptops with limited RAM.

When to Use This Model

  • Specific Alignment Needs: Ideal for applications requiring an AI with a defined behavioral and ethical framework, as dictated by the Beneth DOCTRINE.
  • General Purpose & Coding Tasks: Suitable for tasks involving general knowledge queries, logical problem-solving, and code generation or assistance in Python, SQL, and JavaScript.
  • Resource-Constrained Environments: The GGUF quantization allows for efficient inference on devices with limited computational resources, broadening its accessibility.