Google Gemini

In December of 2023, Google publicly released their own Large Language Model (LLM) to compete with OpenAI's ChatGPT. Their service called Gemini, is available at https://gemini.google.com/.

Google's release of Gemini has faced controversy, including an early version of their image generation that was widely criticized for being inaccurate when generating images of historical figures.¹

Google Gemini has the following models available for use through the Gemini API:

• Gemini 1.5 Flash: A fast multimodal model that accepts audio, images, video and text and outputs text. It has a text token limit of 1,048,576 with an output token limit of 8,192. It has both a free and pay-as-you-go tiers based on usage.²
• Gemini 1.5 Pro: A mid-size multimodal model for a wide-range of reasoning task that accepts audio, images, video, text and outputs text. It has a text token limit of 2,097,152 with an output token limit of 8,192. It has both a free and pay-as-you-go tiers based on usage.³
• Gemini 1.0 Pro: A Natural Language Processing(NLP) model for text, code chat, and code generation that accepts text and outputs text. It has both a free and pay-as-you-go tiers based on usage.⁴

Using Gemini

Gemini can be used through a web interface or programmatically through the Gemini API. Google provides Python, Go, Node.js, Web, Dart, Swift, and Android libraries that wrap the API calls.⁵

Workshop Exercise

Create a Github issue for a DSPy model signature of a FOLIO Circulation Record