Google Launches Gemma 4 12B Multimodal AI Model Without Encoders

What Happened

Google has unveiled Gemma 4 12B, a versatile AI model capable of processing both text and image data without relying on traditional encoders. Gemma 4 12B is part of Google’s ongoing effort to advance multimodal AI and aims to make it easier for developers to implement unified text and vision functionalities in their products. This encoder-free approach simplifies architecture and boosts performance on a range of multimodal tasks. The model is available for research and commercial use, further extending Google’s AI ecosystem.

Why It Matters

Gemma 4 12B reflects a significant shift toward unified AI solutions that reduce system complexity and improve efficiency in handling multimodal data. Google’s latest release is expected to spur innovation and facilitate broader adoption of advanced AI tools. Read more in our AI News Hub