Google Releases Gemini

Google introduced Gemini in December 2023 as a new multimodal model family with Nano, Pro, and Ultra variants. Rather than treating images, text, and other modalities only as separate systems connected afterward, Google emphasized multimodal training and reasoning as a core design goal.

Gemini Nano targeted on-device use, Pro powered services including Bard, and Ultra was presented as the largest version for more demanding tasks. The family could process combinations of text, images, audio, and code depending on the product.

The launch was strategically important after the rapid rise of ChatGPT and GPT-4. Google was bringing DeepMind and Google Brain research under one model program while trying to deploy the results across Search, Android, cloud services, and consumer applications.