Introducing Gemma 4 12B: a unified, encoder-free multimodal model

Key points

• Today, we are introducing Gemma 4 12B, our latest model designed to bring agentic multimodal intelligence directly to laptops.1
• Here’s an overview of what makes Gemma 4 12B unique:1
• Drafter-ready: Gemma 4 12B comes equipped with Multi-Token Prediction (MTP) drafters to reduce latency.1
• Here is how Gemma 4 12B processes multimodal inputs natively:1

Sources (1)

• [1]Introducing Gemma 4 12B: a unified, encoder-free multimodal model
  Google DeepMind Blog · Jun 9, 02:10 PM

  Introducing Gemma 4 12B: a unified, encoder-free multimodal model

  Today, we are introducing Gemma 4 12B, our latest model designed to bring agentic multimodal intelligence directly to laptops.

Extractive summary: sentences quoted from the sources.