Gemini (family)

Google DeepMind's flagship LLM family — Gemini 1.0, 1.5 (1M context), 2.0/2.5 (multimodal, reasoning) — competes directly with GPT and Claude.

Gemini is Google DeepMind's flagship multimodal LLM family. The line: Gemini 1.0 (Dec 2023, the first natively-multimodal frontier model), Gemini 1.5 (1M-token context, retrieval over long docs), Gemini 2.0 (with native image and audio output), Gemini 2.5 Pro and Flash (with reasoning). It matters because Gemini is the third major player in the frontier LLM market alongside GPT and Claude. Google's distinctive bets: native multimodality (audio, video, image, text trained together rather than bolted on), the longest production context windows (1M and 2M tokens), and tight integration with Google products — Search, Maps, Gmail, Docs, YouTube. For tasks involving image+text or long video understanding, Gemini is often the strongest available option. The family includes Gemini Flash (fastest tier, very cheap) and Gemini Ultra (the most capable but lower-volume tier). Gemini 2.5 Pro became the default in many Google products in 2025. Gemini is accessible via Google AI Studio, the Vertex AI platform on Google Cloud, the Gemini app (consumer), and tightly embedded in Google Search (AI Overviews), Workspace, and Bard's successor product. Like GPT and Claude, Gemini is closed source — Google's open-weight family is the separately-named Gemma. Related: Google DeepMind, Gemma, multi-modal, long-context.