Gemma 3 12B

Gemma 3 12B is a lightweight model from Google's Gemma family, designed to balance capability with efficiency for moderate-scale applications. As part of Google's third-generation Gemma series, this 12-billion parameter model sits in the lightweight tier, making it suitable for applications that need reasonable performance without the computational overhead of larger flagship models. The model supports both text and image inputs with a 131,072 token context window, allowing it to process substantial documents or maintain extended conversations while incorporating visual information. Performance benchmarks show an output speed of 28.89 tokens per second with a time to first token of approximately 27 seconds, indicating steady generation once processing begins. Gemma 3 12B serves applications requiring multimodal understanding at scale, such as document analysis with visual elements, content moderation, or customer service scenarios where both text and image processing are needed but computational resources are constrained compared to flagship model deployments.