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.