Distilled - Gemma 3 12B

This model is a Gemma 3 12B variant distilled from xAI’s Grok 3, with reasoning. It was fine-tuned to emulate Grok’s depth and structured clarity, particularly in tasks involving complex thought, such as problem-solving, coding, and mathematics.

The model was trained on:

This dataset consists of 100 high-quality Grok 3 completions with reasoning responding to deep questions, solving math problems, and writing or analyzing code. The aim was to distill Grok’s analytical approach and technical versatility into a smaller, accessible model.

This Gemma model was trained 2x faster with Unsloth and Huggingface's TRL library.