Google DeepMind·Gemini

Gemini 3.8 Flash

The stable Gemini Flash upgrade for long-horizon software engineering, autonomous agents, and complex enterprise workflows, with multimodal input, 1M context, and 65K output.

CurrentLanguageReasoningCodingAgentMultimodalWeights not released

MODEL NUMBER095

First public date: 2026-09-02
Organization: Google DeepMind
Family: Gemini

Specifications

Only published specifications are shown.

Context: 1,048,576 tokens
Maximum input: 1,048,576 tokens↗
Maximum output: 65,536 tokens↗
Input modalities: Text · Image · Audio · Video
Output modalities: Text
Open weights: No
Status: Current

Overview

Release dates

September 2, 2026 — Gemini 3.8 Flash launched with a 1M input context, 65K output limit, multimodal inputs, and thinking capabilities. ↗

Why it mattered

It extends Flash-level speed and efficiency to longer engineering and agentic tasks while retaining full tool use and thinking levels.

Specifications alongside predecessor

Compared with Gemini 3.7 Flash ↗

PreviousGemini 3.7 Flash— · 1M ctx

CurrentGemini 3.8 Flash— · 1.05M ctx

Architecture & capabilities

Model capabilities

CodingAgentTool useLong context

Connected tools

Computer useGrounding

API features

Thinking levels

Release history

2026-09-02General release

Gemini 3.8 Flash launched with a 1M input context, 65K output limit, multimodal inputs, and thinking capabilities.

View source

Related models

Select a card to view the related model.

PredecessorGemini 3.7 FlashView details →VariantGemini 3.8 Flash CyberFairwind limited accessView details →

Sources

Gemini 3.8 FlashGoogle · Published 2026-09-02

Official · Verified 02

Gemini 3.8 Flash model limitsGoogle · Checked 2026-09-07

Official · Verified

Suggest a correction

Copy this template, add the field, proposed value, and primary source, and share it with the archive maintainer.

Model ID: gemini-3-8-flash
Page: https://llmlineage.com/en/models/gemini-3-8-flash
Field:
Proposed value:
Primary source URL:
Reason: