---
title: "Gemini 3.8 Flash in the EU: API with EU data residency"
description: "Use Google's Gemini 3.8 Flash on inference.eu: a fast model with 1M context, requests processed in EU data centres and nothing used for training."
canonical: "https://inference.eu/models/gemini-3-8-flash/"
language: "en"
---

Chat · Google DeepMind

# Gemini 3.8 Flash in the EU, processed in European data centres

Gemini 3.8 Flash is Google's most capable Flash model, released in September 2026 for coding, agents and business workflows at Flash speed, with a context window of about one million tokens. On inference.eu you call it through an OpenAI-compatible API, processed in EU data centres and never used for training.

[Contact sales](https://inference.eu/contact/)[Example request](#example)

1.  01Context window: 1,048,576 tokens
2.  02EU data centres
3.  03Proprietary

## What Gemini 3.8 Flash is

Gemini 3.8 Flash is a closed model from Google DeepMind, released on 2 September 2026 and built on Gemini 3.7 Flash, which it has since replaced. Google describes it as its most intelligent Flash model. It accepts text, images, video, audio and PDF documents and produces text.

It takes up to 1,048,576 input tokens and can write up to 65,536 tokens in one response. Thinking is adjustable in three levels, low, medium and high, with medium as Google's default; the minimal level of some earlier Flash models is not supported. Google does not publish the weights or the parameter count.

## What it does well, and where it falls short

Google built 3.8 Flash for long-horizon software engineering, autonomous agents and complex enterprise workflows, while keeping the speed and cost profile of the Flash line. It is a good default for chat assistants, document question answering, extraction from long files and agents that call many tools.

Google notes that on long, complex tasks the model deliberately takes smaller reasoning steps, calls tools repeatedly and checks its work, so it can use more tokens than you might expect. Set the thinking level to low for simple requests. For the hardest problems, such as large code migrations or deep research, Gemini 4 Argon goes further.

## Calling Gemini 3.8 Flash through the API

inference.eu speaks the OpenAI request format, so the OpenAI SDKs for Python and JavaScript, LangChain, LlamaIndex and the Vercel AI SDK work as they are. Set the base URL to the inference.eu API, use your inference.eu key and set the model to `gemini-3.8-flash`. Tools that use the Anthropic Messages format can call the same model there.

Send images as `image_url` content parts alongside your text. Tool calls, JSON output and streaming work as usual. If you plan to send audio, video or PDF files, get in touch first so we can confirm the request format for your account.

## Gemini with EU data residency

When you call Gemini 3.8 Flash on inference.eu, your prompts, files and the model's responses are processed in data centres in the European Union. They are used to answer your request and are never used to train any model.

We sign one data processing agreement under Article 28 of the GDPR with every customer and provide our list of subprocessors, so your data protection officer can see who handles the data and where. The same agreement covers Gemini, Claude and every open model on inference.eu, all billed in euros on one invoice.

## Gemini 3.8 Flash

Maker

Google DeepMind

Made in

United States

Type

Chat

Context window

1,048,576 tokens

Licence

[Proprietary](https://ai.google.dev/gemini-api/terms)

Released

September 2026

Model ID

gemini-3.8-flash

Processing

EU data centres

Training on your data

Never

Source

[ai.google.dev](https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash)

gemini-3.8-flash · Python

```
import os
from openai import OpenAI

client = OpenAI(base_url="https://api.inference.eu/v1", api_key=os.environ["INFERENCE_EU_API_KEY"])

response = client.chat.completions.create(
    model="gemini-3.8-flash",
    messages=[{"role": "user", "content": "Summarise this contract in five bullet points."}],
)
print(response.choices[0].message.content)
```

## What teams use it for

-   1

    ### Company assistants

    Answer staff questions over long policies, manuals and reports that fit into a single request thanks to the one-million-token window.

-   2

    ### Agents with many tool calls

    Run agents that search, call your APIs and check their own results over many steps, with every request processed in the EU.

-   3

    ### Document extraction

    Pull fields, tables and figures out of long reports and scanned pages into JSON that your systems can process.

## Gemini 3.8 Flash: common questions

The questions buyers, developers and legal teams ask us most. Something missing? Ask us directly.

Yes. On inference.eu, Gemini 3.8 Flash is served from EU-only endpoints and your requests are processed in data centres in the European Union. You get one data processing agreement for all models and one invoice in euros.

Whether your use is lawful depends on what you process and why, not on the model alone. On inference.eu, requests are processed in EU data centres, are not used for training, and are covered by a data processing agreement under Article 28 of the GDPR with a published subprocessor list. That gives you the contractual basis to use Gemini 3.8 Flash with personal data.

No. What you send to inference.eu is used only to produce your response. It is not used to train or fine-tune Gemini or any other model.

Gemini 3.8 Flash accepts up to 1,048,576 input tokens and writes up to 65,536 tokens per response. That is enough for long contracts, full codebases of moderate size or several reports in one request.

Google lists text, images, video, audio and PDF as inputs, and text as output. On inference.eu, text and images work through the standard OpenAI message format. Contact us before building on audio, video or PDF input so we can confirm the format for your account.

There are three thinking levels: low, medium and high, with medium as the default. The minimal level is not supported and returns an error. Use low for quick answers and high for multi-step reasoning.

Gemini 3.8 Flash is based on Gemini 3.7 Flash and replaces it; Google has deprecated 3.7 Flash. Change the model string to \`gemini-3.8-flash\`. Expect better results on coding and agent tasks, and check token use on long tasks, where 3.8 Flash may work in more steps.

Start with Gemini 3.8 Flash for most work, especially at high volume. Move to Gemini 4 Argon for the hardest tasks, such as long agent runs, complex code changes and in-depth analysis, where extra reasoning pays off.

No. Gemini 3.8 Flash is a closed model and Google does not publish the weights. Use is governed by Google's terms. Google's open models are the Gemma family; Gemma 4 31B and Gemma 4 26B A4B are also available on inference.eu.

Yes. Change the base URL to the inference.eu API, use your inference.eu key and set the model to \`gemini-3.8-flash\`. Your messages, tools and streaming code stay the same.

[Ask a question](https://inference.eu/contact/)

## Related

[All models](https://inference.eu/models/#directory)

-   [Gemini 4 ArgonUse Google's Gemini 4 Argon on inference.eu: requests processed in EU data centres, one data processing agreement, and nothing used for training.Reasoning](https://inference.eu/models/gemini-4-argon/)
-   [Gemma 4 31BRun Google's Gemma 4 31B on inference.eu: an open multimodal model with 256K context behind an OpenAI-compatible API, processed in EU data centres.Vision](https://inference.eu/models/gemma-4-31b/)
-   [Gemma 4 26B A4BRun Google's Gemma 4 26B A4B on inference.eu: a fast mixture-of-experts model with image input and 256K context, processed in EU data centres.Vision](https://inference.eu/models/gemma-4-26b-a4b/)

## Move your AI workloads to EU infrastructure

Tell us which models you use and roughly how much traffic you have. We reply within one working day with pricing and next steps.

-   Reply within one working day
-   Pricing for your volume
-   DPA ready to sign