Back to Gemini 3
Google DeepMind /

Gemini 3.6 Flash
Gemini 3.6 FlashFeatured

Closed SourceChat & ReasoningtextimageaudiovideoUpdated July 21, 2026

Gemini 3.6 Flash: Google's Fastest Frontier Model

Model Overview

Gemini 3.6 Flash is Google DeepMind's fastest frontier-class model, released on July 21, 2026 alongside Gemini 3.5 Flash-Lite and Gemini 3.5 Flash Cyber. It is ranked #1 fastest model among 186 models tested by Artificial Analysis at 275.5 output tokens per second, while scoring 50 on the Artificial Analysis Intelligence Index — well above the 31-point average for comparable models. With native multimodal input (text, image, audio, video), a 1M token context window, and built-in hybrid reasoning, Gemini 3.6 Flash occupies the sweet spot between frontier intelligence and deployment speed.


✨ Key Features

Specification Table
FeatureDescription
#1 Fastest Model275.5 output tokens/second — ranked fastest of 186 models on Artificial Analysis
Intelligence Index: 50Well above the 31-point average for comparable models
Hybrid ReasoningBuilt-in thinking mode for complex multi-step reasoning tasks
Full Multimodal InputAccepts text, image, audio, and video in a single context
1M Token ContextUltra-long context window for document analysis and complex tasks
90% Cache DiscountCached input tokens billed at $0.15/1M (vs. $1.50 standard)

📊 Benchmarks

Specification Table
BenchmarkScoreSource
Artificial Analysis Intelligence Index50Artificial Analysis
Output Speed275.5 tok/s (#1 / 186)Artificial Analysis
Intelligence Rank#21 / 186Artificial Analysis

💰 Pricing

Specification Table
TierPrice per 1M tokens
Input$1.50
Output$7.50
Cache Hit$0.15 (90% discount)

🔗 Resources

Specification Table
ResourceLink
Blog Announcementblog.google
Model Card (PDF)Gemini 3.6 Flash Model Card
Try it in AI Studioaistudio.google.com
Artificial Analysisartificialanalysis.ai/models/gemini-3-6-flash

📜 License & Access

Proprietary — Available via Google AI Studio and the Gemini API.

Key Features

Ranked #1 fastest model among 186 models at 275.5 output tokens/second (Artificial Analysis)

Feature 01

Intelligence Index score of 50, above the 31-point average for comparable models

Feature 02

Built-in hybrid reasoning mode with extended thinking capabilities

Feature 03

Native multimodal input: text, image, audio, and video in a single context

Feature 04

1M token context window — 90% cache pricing discount at $0.15/1M cached tokens

Feature 05

You might also want to compare

Verified Sources

Tags

reasoningmultimodalfastlong-context

Model Specs

closed-source

Parameters

Undisclosed

Context Window

1M tokens

License

Proprietary

Deployment

api-only

Resources & Links

Lineage

Model Family

Part of the Gemini 3 family

Only release in this line currently tracked.

Predecessor

Gemini 3.5 Flash

Curator Notes

Released July 21, 2026 alongside 3.5 Flash-Lite and 3.5 Flash Cyber. Ranked #1 fastest model on Artificial Analysis. Model card PDF available at storage.googleapis.com.

Compare Specs

Compare parameters, context windows, modalities, and benchmark scores of this model side-by-side with others.

Compare Model