---
title: "Qwen3.6 27B API hosted in the EU"
description: "Run Qwen3.6 27B on inference.eu: a dense open model for agentic coding with image input and 256K context, OpenAI-compatible and processed in the EU."
canonical: "https://inference.eu/models/qwen3-6-27b/"
language: "en"
---

Code · Qwen (Alibaba Cloud)

# Run Qwen3.6 27B in the EU through an OpenAI-compatible API

Qwen3.6 27B is the dense member of Alibaba's Qwen3.6 family, released in April 2026 with a focus on agentic coding. It reads images, has a thinking mode and handles 256K tokens of context. On inference.eu you call it with the OpenAI or Anthropic request format, processed in EU data centres and never used for training.

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

1.  0127B
2.  02EU data centres
3.  03Apache 2.0

## What Qwen3.6 27B is

Qwen3.6 27B is a 27-billion-parameter language model with a vision encoder, released under Apache 2.0. It is a dense model, so all parameters work on every token, but its 64 layers mix Gated DeltaNet linear-attention blocks with gated attention blocks in a three-to-one pattern, which keeps long contexts efficient.

Its native context is 262,144 tokens, and Qwen describes ways to extend it to about one million. Qwen built the 3.6 release on community feedback after Qwen3.5, with priority on stability and practical coding: front-end workflows, repository-level reasoning and an option to keep reasoning from earlier turns in the conversation, called thinking preservation.

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

Qwen3.6 27B is a strong choice for coding assistants and agents at a size that is easy to serve. It handles front-end tasks, multi-file changes and tool use well, reads screenshots and diagrams, and its thinking mode helps with debugging and planning. The non-thinking mode gives quick answers when you do not need a reasoning trace.

It is a mid-sized model: on the hardest long-horizon tasks, GLM-5.3, Kimi K3 or DeepSeek V4 go further. Qwen advises keeping at least 128K of context available to preserve its thinking quality, which matters if you trim histories aggressively. Its successor, Qwen3.8 27B, is the newer choice for most new projects.

## Calling Qwen3.6 27B through the API

inference.eu uses the OpenAI request format, so the OpenAI SDKs and coding tools such as Cline, OpenHands or Claude Code through our Anthropic-compatible endpoint work as they are. Set the base URL to the inference.eu API, use your inference.eu key and set the model to `qwen3.6-27b`.

Thinking is on by default. To get a direct answer, pass `enable_thinking: false` in `chat_template_kwargs` through `extra_body`; to keep earlier reasoning in context, pass `preserve_thinking: true`. Qwen recommends `temperature` 1.0 for general thinking tasks and 0.6 for precise coding. Send images as `image_url` content parts.

## Data residency for Qwen3.6 27B

Qwen comes from Alibaba Cloud in China, but on inference.eu Qwen3.6 27B runs on infrastructure in the European Union. Prompts, source code, images and responses are processed in EU data centres, are not sent to Alibaba, and are never used to train any model.

Every customer receives a data processing agreement under Article 28 of the GDPR and our subprocessor list. Billing is in euros. The Apache 2.0 licence allows commercial use, modification and redistribution without user limits or revenue thresholds, so there is no separate agreement with Alibaba to negotiate.

## Qwen3.6 27B

Maker

Qwen (Alibaba Cloud)

Made in

China

Type

Code

Parameters

27B

Context window

262,144 tokens

Licence

[Apache 2.0](https://huggingface.co/Qwen/Qwen3.6-27B/blob/main/LICENSE)

Released

April 2026

Model ID

qwen3.6-27b

Processing

EU data centres

Training on your data

Never

Source

[huggingface.co](https://huggingface.co/Qwen/Qwen3.6-27B)

qwen3.6-27b · 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="qwen3.6-27b",
    messages=[{"role": "user", "content": "Summarise this contract in five bullet points."}],
)
print(response.choices[0].message.content)
```

## What teams use it for

-   1

    ### IDE coding assistant

    Answer code questions, write and refactor functions and explain errors, at a size that responds quickly.

-   2

    ### Front-end from designs

    Turn a screenshot or mock-up into components and styles, then iterate on them with feedback.

-   3

    ### Repository-level agents

    Let an agent read several files, plan a change and apply it, with code processed in the EU.

## Qwen3.6 27B: common questions

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

Compliance depends on how you process data. On inference.eu, Qwen3.6 27B requests are processed in EU data centres, not used for training and covered by a data processing agreement under Article 28 GDPR with a published subprocessor list. Nothing is sent to Alibaba or to China.

The native context is 262,144 tokens. Qwen describes methods to extend it to about one million tokens. Qwen also advises keeping at least 128K tokens of context available so the model's thinking quality is preserved.

Yes. Qwen3.6 27B is released under the Apache 2.0 licence, which allows commercial use, modification and redistribution without fees or user limits. Keep the licence notice only if you redistribute the weights yourself.

Thinking is on by default. Pass \`enable\_thinking: false\` inside \`chat\_template\_kwargs\` in the request's \`extra\_body\` to get a direct answer. Qwen recommends \`temperature\` 0.7 and \`top\_p\` 0.8 for this non-thinking mode.

It is an option introduced in Qwen3.6 that keeps the reasoning from earlier assistant messages in the context. In iterative coding sessions this saves the model from re-deriving the same conclusions. Enable it with \`preserve\_thinking: true\` in \`chat\_template\_kwargs\`.

Yes. It includes a vision encoder, so it can read screenshots, diagrams and scanned pages together with text. It produces text only. Send images using the standard OpenAI \`image\_url\` content format.

Qwen3.8 27B is the newer model of the same size, with gains in coding, agent tasks and video understanding plus a \`reasoning\_effort\` setting. For new projects it is usually the better default. Qwen3.6 27B suits teams that have already tested and tuned prompts on it.

The 27B model is dense and generally gives stronger answers. The 35B A3B is a mixture-of-experts model with about 3 billion active parameters, which makes it much lighter per token. Choose 27B for quality, 35B A3B for speed and volume.

On inference.eu the open weights run in EU data centres, so your code and prompts do not go to Alibaba or China. The model's training still reflects its maker's choices, including on politically sensitive topics. For coding this rarely matters, but test it on your own tasks.

Change the base URL of your OpenAI client to the inference.eu API, use your inference.eu key and set the model to \`qwen3.6-27b\`. Messages, images, tools and streaming keep working. Add \`chat\_template\_kwargs\` options if you want to control thinking.

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

## Related

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

-   [Qwen3.6 35B A3BRun Qwen3.6 35B A3B on inference.eu: a fast mixture-of-experts model with 3B active parameters, image input and 256K context, processed in the EU.Code](https://inference.eu/models/qwen3-6-35b-a3b/)
-   [Qwen3.8 27BRun Qwen3.8 27B on inference.eu: a dense open model with image and video input, adjustable reasoning and 256K context, OpenAI-compatible, processed in the EU.Reasoning](https://inference.eu/models/qwen3-8-27b/)
-   [Devstral Small 2Run Devstral Small 2, Mistral AI's open 24B coding agent model, on inference.eu with OpenAI and Anthropic APIs and processing in EU data centres.Code](https://inference.eu/models/devstral-small-2/)
-   [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/)

## 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