unsloth/gemma-3-27b-it-qat

The unsloth/gemma-3-27b-it-qat model is a 27 billion parameter instruction-tuned variant of Google DeepMind's multimodal Gemma 3 family, built from the same research as Gemini models. This version utilizes Quantization Aware Training (QAT) to maintain bfloat16 quality while significantly reducing memory requirements, making it suitable for resource-constrained environments. It supports a 128K token context window, handles both text and image inputs, and excels at text generation, image understanding, and reasoning tasks across over 140 languages.