Gemini 3.6 Flash

Most token-efficient Flash model: 17% fewer output tokens and fewer reasoning steps than 3.5 Flash for multi-step workflows

Input

$1.50

per 1M tokens

Output

$7.50

per 1M tokens

Cached

$0.15

per 1M tokens

Context

1.0M tokens

Max Output

66K tokens

Released

2026-07-21

Features

[thinking-model][17%-fewer-output-tokens-vs-3.5-flash][1m-tokens-context][fewer-reasoning-steps-for-agentic-tasks][prompt-caching]

Technical Specifications

The Lab

location: London, UK, and Mountain View, California, USAfounded: 2023backed by: Alphabet Inc. (subsidiary)

Formed in 2023 by merging DeepMind (founded 2010) and Google Brain (founded 2011) into a single Alphabet subsidiary headquartered in London and Mountain View. The lab integrates frontier AI research with Google product ecosystems, producing natively multimodal models, open-weights Gemma variants, and scientific models like AlphaFold. Its Gemini family anchors Google's AI product stack across search, cloud, and consumer applications.

more about Google DeepMind →

Model ID: gemini-3.6-flash