Google DeepMind's Veo video model, now presented as Veo 3.1: cinematic video with native audio, plus 1080p and 4K output on the official model page.

Google's source-grounded research notebook, now presented as Gemini Notebook at notebook.google. Analyzes your files and turns them into audio, video, and more.

Google's agent-first IDE at antigravity.google. Public preview remains free. Do not treat a Windsurf acquisition as a first-party fact.

Google's agent SDK in Python, Java, Go, and TypeScript. Docs now live at google.github.io/adk-docs. A2A is a separate protocol repo.

Nano Banana is Google's Gemini image generation and editing model, led by Nano Banana 2, with a Pro option in the Gemini app.

A2A is an open agent-to-agent communication protocol under the Linux Foundation, built on Agent Cards, tasks, and messages and complementing MCP.

Lightweight multilingual text embedding model from Google DeepMind, optimized for on-device AI with <200MB RAM usage.

Spotify's vendor-neutral agentic development environment. One place to launch, manage, and switch between Claude Code, Gemini CLI, and Codex sessions, with p.

Gemini 2.0 Flash Thinking was Google's experimental reasoning model, released in December 2024 and scheduled for shutdown on December 2, 2025.

Gemini 1.5 Flash-8B is Google's smaller, cheaper Gemini 1.5 Flash variant, made production-ready on October 3, 2024 at $0.0375 per 1M input tokens.

Gemini 2.0 Flash was shut down in the Gemini API on June 1, 2026. It had a 1,048,576-token input limit, multimodal input and a knowledge cutoff of August 2024.

Google's latest frontier model delivering breakthrough intelligence at unprecedented speed and cost efficiency.

Gemini 3 Pro Preview (gemini-3-pro-preview), released Nov 18, 2025, is shut down; Google names gemini-3.1-pro-preview as the replacement.

Gemini 1.5 Flash (gemini-1.5-flash) was Google's fast multimodal Gemini API model, released in 2024 and shut down on September 29, 2025.

Gemini 1.5 Pro, announced February 15, 2024, was a mixture-of-experts multimodal model with up to 1M tokens of context. Its API was shut down in 2025.

Gemma 2 27B is Google's open-weights 27B decoder-only text model, trained on 13T tokens, in pretrained and instruction-tuned variants under the Gemma license.

Gemma 2 9B is Google's open-weights 9B text model released June 27, 2024, trained on 8T tokens; Gemma 4 is now the latest Gemma generation.

Chrome built-in Gemini. Auto browse preview is U.S. AI Pro/Ultra, not free-for-all.

Run Claude, Codex, Gemini and other AI agents in parallel in a split-screen terminal, with persistent context and low token overhead.

BAAI bge-reranker-v2.5-gemma2-lightweight is a roughly 9B multilingual reranker on Gemma 2 9B with token compression and layer cutoffs.

A third-party Claude Code skill that drives NotebookLM through a local browser; it works only with local Claude Code, not the web UI sandbox.

Citation skill that searches Google Scholar and PubMed, turns DOIs, PMIDs, and arXiv IDs into BibTeX, and validates bibliographies before submission.

Google's open-source orchestrator for agentic workloads. Declare Tasks, Workspaces, and Models in YAML and AX sandboxes and runs them on Kubernetes.