ModelRadar
/

Gemini 3.6 Flash

Google DeepMind🇺🇸 USFrontier· Jul 21, 2026

TextReasoningVisionAPI

Gemini 3.6 Flash consumes 17 % fewer output tokens than its predecessor and completes multi-step workflows with fewer reasoning steps and tool calls. It targets coding, knowledge work, and agentic automation at lower cost.

Summary of the model card

Context
1M tokens
Price / 1M tok
$0.38 in · $1.88 out
Modalities
text, image, video, document, audio → text

2 variants

Standard1M$0.75/$3.75OR
Batch1M$0.38/$1.88OR

Sources

OpenRouterBlog
Deutsch

Das Modell verbraucht 17 % weniger Ausgabetokens als der Vorgaenger und bearbeitet Mehrschrittaufgaeben mit weniger推理-Schritten und Tool Calls. Es richtet sich an Coding, Wissensarbeit und agente Automatisierung zu niedrigeren Kosten.

Provenance

Detected
Sep 27, 2026
First source
openrouter
Description
Summary of the model card
Editorially reviewed
—