Google Unveils Open AI Model Gemma

On February 21, Google released the open AI model Gemma for responsible artificial intelligence development.

It is available in two versions — Gemma 2B and Gemma 7B, both offered under a commercial license. These models are lightweight enough to operate not only in Google’s cloud but also on a laptop or desktop computer.

Google claims that in tests of massive multitask language understanding (MMLU), they outperform Mistral 7B and Llama 13B.

Users can access the models through the data analysis platform Kaggle, Colab notebooks, or Google Cloud. A more powerful version, Gemma 1.5 Pro, is in development, with its release date yet to be announced.

Back in December, Google released a global update for the chatbot Bard, aimed at significantly enhancing the capabilities of the neural network.