Google Unveils New Open-Al Model ‘Gemma’

Google has unveiled a new artificial intelligence (AI) model, ‘Gemma’, which is a free open-weight model built on technology similar to the more powerful but closed Germni model.

Gemma models share technical and infrastructure components with Germni, Google’s largest and most capable AI model widely available today.

This enables Gemma 2B and 7B to achieve best-in-class performance for their sizes compared to other open models. And Gemma models are capable of running directly on a developer laptop or desktop computer. Notably, Gemma surpasses significantly larger models on key benchmarks while adhering to our rigorous standards for safe and responsible outputs.

Gemma models run across popular device types, including laptop, desktop, IoT, mobile and cloud, enabling broadly accessible AI capabilities.

However, Google hasn’t made Gemma entirely open source, retaining control over terms of use and ownership. While some experts caution against potential misuse of open-source AI, others advocate for its ability to expand participation and benefits across a broader community.

Also Read: Google unveils new standalone AI image generator

Google chose not to release its larger premier Gemini models alongside the announcement, unlike Gemma. Google stated that Gemma models come in sizes of two billion or seven billion parameters, representing the number of different values considered by the algorithm to generate output.

Meta’s Llama 2 models vary in size from seven to 70 billion parameters. The specific size of Google’s largest Gemini models remains undisclosed. For context, OpenAI’s GPT-3 model, unveiled in 2020, comprised 175 billion parameters.

Nvidia, the chipmaker, announced on Wednesday that it collaborated with Google to optimize Gemma models for its chips.

Additionally, Nvidia stated its intention to integrate its chatbot software, currently in development for Windows PCs, with Gemma in the near future.