choonok/VetJarvis-1.1-4B-Instruct
VetJarvis-1.1-4B-Instruct by choonok is a 4.5 billion parameter instruction-tuned large language model based on the Qwen3.5-4B architecture, featuring a 32768 token context length. This upgraded version focuses on enhanced reasoning capabilities, achieving logical reasoning performance comparable to GPT-5.4-mini through an improved post-training process. It is specifically designed as a clinical-support AI assistant for veterinarians, excelling in tasks requiring detailed reasoning.
Loading preview...
VetJarvis 1.1-4B-Instruct Overview
VetJarvis 1.1-4B-Instruct is an enhanced 4.5 billion parameter language model developed by choonok, built upon the Qwen3.5-4B architecture. This iteration significantly improves upon its predecessor, VetJarvis 1.0, by focusing on advanced reasoning capabilities through a refined post-training pipeline. It is specifically designed to function as a clinical-support AI assistant for veterinarians.
Key Improvements & Capabilities
- Enhanced Reasoning Performance: Version 1.1 delivers meaningful performance gains, particularly when the
thinkoption is enabled, addressing limitations of the previous release. - GPT-5.4-mini Class Reasoning: Through its optimized post-training, VetJarvis 1.1 achieves logical reasoning performance on par with models like GPT-5.4-mini.
- Optimized for Clinical Support: The model is fine-tuned to assist veterinarians, making it suitable for specialized medical queries and protocols.
- High Context Length: Supports a substantial context length of up to 32,768 tokens, with recommendations for up to 262,144 tokens.
Recommended Usage
- Temperature: 0.8
- Top-p: 0.9
- max_new_tokens: 32,768
- enable_thinking:
True(highly recommended for optimal performance)
This model is ideal for applications requiring robust reasoning in a veterinary context, offering a powerful tool for clinical decision support.