Google Gemini API

Google's API for Gemini models, called through the Interactions API with the google-genai SDK. Supports text, vision, video understanding, structured JSON output, tool use, Google Search grounding, live voice, TTS and transcription.

Website Docs Changelog Compare with others

Google Gemini API at a glance

From official pages, checked Nov 15, 2026. Tap the arrow for each source.

Free tier: Yes
Free of charge on Flash and Flash-Lite models and Gemma 4; free-tier content is used to improve products.
Starting price: $0.10 per 1M input tokens
Gemini 2.5 Flash-Lite standard text input, $0.40 output; cheapest paid standard rate on the page.
Size limit: 1M token context
Gemini 3.8 Flash: 1,048,576 input tokens, 65,536 output tokens.
Official SDKs: PythonJavaScript / TypeScript
Trains on your data: No
Paid tier prompts and responses are not used to improve products; free tier data is.
EU hosting: Yes
EU multi-region endpoints via Google Cloud (Gemini Enterprise Agent Platform), not AI Studio.

What you can build with Google Gemini API

Google Gemini APIPro archiveSnippetPaste

Describe images and answer questions about them with Gemini vision in Python

A Python script that sends a local image to gemini-4-hydrogen and prints a caption or the answer to any question about it.

Free tier: YesFrom $0.10 per 1M input tokens

Vision

Recent updatesAll updates

Nov 12, 2026
Gemini 4 Hydrogen is generally available: $1.50 / $7.50, a free tier and a new xhigh thinking level (November 12, 2026)
gemini-4-hydrogen is GA with a free tier. Thinking defaults to high, minimal is rejected, and explicit caching is not supported.
Nov 11, 2026
Gemini Decision and the Decisions API: yes/no, choice and score answers with zero output tokens (November 11, 2026)
New gemini-decision-1 model returns yes/no, choice and score answers, billed on input only at $0.10 per 1M tokens on the paid tier.
Nov 10, 2026
Gemini 4 Carbon is generally available: $2 / $10 until November 30, no temperature or explicit caching (November 10, 2026)
gemini-4-carbon is GA on the paid tier. Prices double Dec 1. Sampling parameters are ignored and explicit caching is not supported.
Nov 6, 2026
gemini-embedding-2-preview is deprecated: calls now route to gemini-embedding-2, no re-indexing needed (November 6, 2026)
Calls to gemini-embedding-2-preview now go to gemini-embedding-2. Change the model string. Stored vectors still match.
Nov 6, 2026
Google Gen AI SDK 3.0 for JavaScript and Python defaults to API v1, and the JavaScript SDK needs Node.js 22 (November 6, 2026)
@google/genai and google-genai 3.0.0 call API v1 by default, so -latest aliases break. JS needs Node 22. Python switches to httpx2.
Nov 5, 2026
Gemini Omni 1.2 Flash is generally available with cheaper video output (November 5, 2026)
gemini-omni-1.2-flash makes 15-second videos with audio for $12 per 1M video tokens, down from $17.50. Omni 1.1 Flash reroutes on Dec 3.
Nov 5, 2026
Gemini tools can now return Office files, audio and video to the model (November 5, 2026)
Gemini function responses now take Office files, audio, video and Files API URIs. PPT, ODP and ODS files are accepted as input too.
Nov 4, 2026
Gemini API adds voice remixing: change a voice with plain words and save it as a new one (November 4, 2026)
Gemini 3.8 TTS can remix a voice from plain-English instructions and store it as a new voice ID for speech generation.
Nov 2, 2026
Gemini 2.5 Pro and 2.5 Flash deprecated: calls route to 3.8 Flash and 3.5 Flash-Lite from December 1, 2026
From Dec 1, 2026 gemini-2.5-pro calls go to gemini-3.8-flash and gemini-2.5-flash to gemini-3.5-flash-lite. Switch and test first.
Oct 30, 2026
Gemini API Computer Use is now generally available (October 30, 2026)
Google moved the Gemini API Computer Use tool to general availability on Oct 30, 2026, in the Interactions and GenerateContent APIs.
Oct 28, 2026
Google releases the Gemini API CLI (gemini-api) in beta (October 2026)
Google's new gemini-api CLI lets coding agents, scripts and CI call Gemini from the shell: text, media, file analysis and managed agents.
Oct 27, 2026
Gemini API stores Generate Content requests by default for new paid projects (October 26, 2026)
From Oct 26, 2026 new paid-tier Gemini projects log generateContent requests by default. Opt out with store=false or in Logs settings.
Oct 24, 2026
Veo 3.1 preview models and gemini-omni-flash-preview are shut down (October 22, 2026): move to gemini-omni-1.2-flash
Google marks the Veo 3.1 previews and gemini-omni-flash-preview as shut down since Oct 22, 2026. Move to gemini-omni-1.2-flash.
Oct 24, 2026
Gemini TTS previews now shut down Dec 18, 2026, and Live API previews get new dates
Three Gemini TTS previews now shut down Dec 18, 2026, not Nov 17. Native-audio Live preview moves to Mar 31, 2027; 3.1 Live has no date.
Oct 20, 2026
Gemini 4 Argon is generally available on October 13, 2026: $2 input and $10 output per million tokens until November 30
gemini-4-argon is GA. Intro price $2 in, $10 out per 1M tokens through Nov 30, 2026, then $4 and $20. Sampling params are ignored.

Google Gemini API at a glance

What you can build with Google Gemini API

Describe images and answer questions about them with Gemini vision in Python

Extract typed JSON from text with Gemini 4 Hydrogen and Pydantic

Recent updatesAll updates