# Small model efficiency

STILL-3-1.5B-preview is a slow-thinking model enhanced with reinforcement learning technology, achieving 39.33% accuracy on the AIME benchmark

A Python code generation model fine-tuned based on Salesforce/codet5-small, specializing in custom function generation tasks

Intent classification model fine-tuned on the MASSIVE dataset based on Microsoft's Multilingual-MiniLM model, supporting multilingual understanding

Text summarization generation model fine-tuned on the xsum dataset based on T5-small