Gemini

Website: https://deepmind.google/technologies/gemini/

Options: Chatbot, API

Price: multiple prices depending on the model, but they offer free accesss for personal use on some models.

Gemini is the group of LLMs created by Google. These are full multimodal LLMs that can be used for basic text completion or advanced vision and voice tasks. Their most recent releases have strong vision and math capabilities, and they are constantly improving their coding and image features. There are four model families for different purposes:

– Ultra: the largest model for complex tasks

– Pro: general purpose “everything” model

– Flash: lightweight models for speed and efficiency with less complexity

– Nano: intended specifically for on-device tasks. They are heavily focused on optimizing for LLM use on Android devices.

Google is now integrating these LLMs into their search engine to enhance their search results. This new experimental “AI Overview” feature is displayed prominently at the top of most searches.

On December 11th 2024, Google announced a new LLM Agent feature called Deep Research. With just one basic question or prompt, Gemini can repeatedly search the web, learn, and create a complete research report for any topic.