--- title: "DeepSeek V4.1 Flash (1017) API hosted in the EU" description: "Run DeepSeek V4.1 Flash (1017) on inference.eu: a multimodal open reasoning model with 1M context, OpenAI-compatible, processed in the EU, not in China." canonical: "https://inference.eu/models/deepseek-v4-1-flash-1017/" language: "en" --- Reasoning · DeepSeek # Run DeepSeek V4.1 Flash (1017) in the EU through an OpenAI-compatible API DeepSeek V4.1 Flash is DeepSeek's multimodal mixture-of-experts model: it reads images and text, reasons with an adjustable effort from 1 to 100 and handles a one-million-token context. This page covers the September release and its October 1017 update. On inference.eu requests are processed in EU data centres, never in China. [Contact sales](https://inference.eu/contact/)[Example request](#example) 1. 01552B · 16B 2. 02EU data centres 3. 03MIT ## What DeepSeek V4.1 Flash is DeepSeek V4.1 Flash was released in September 2026, followed on 17 October by DeepSeek-V4.1-Flash-1017, a minor update with better reasoning, coding, agent and visual results and shorter outputs. We treat both as one model and serve the 1017 version. It has 552 billion backbone parameters and activates 8 billion parameters per token during prefill and 16 billion during decoding. It uses a new causal encoder-decoder design and a second generation of compressed sparse attention, which DeepSeek says cuts the cache needed per token to about a quarter of V4 Flash. A vision encoder trained from scratch lets it read images natively. The context window is one million tokens, and the weights are MIT-licensed. ## What it does well, and where it falls short V4.1 Flash is aimed at agent work: long coding sessions, multi-step tool use and tasks that mix screenshots, charts and text. Its reasoning effort is a number from 1 to 100, so you can tune cost against accuracy finely instead of picking from three levels. Long inputs are cheap to process because so few parameters are active during prefill. At the top of the effort scale it produces long reasoning traces, so output cost and time rise quickly. It produces text, not images. Like all DeepSeek models, its answers reflect its maker's training choices, including on politically sensitive subjects. For pure text tasks at the lowest cost, DeepSeek V4 Flash or a smaller model may be enough. ## Calling DeepSeek V4.1 Flash through the API inference.eu uses the OpenAI request format, so the OpenAI SDKs and most agent tools work unchanged. Set the base URL to the inference.eu API, use your inference.eu key and set the model to `deepseek-v4.1-flash`. Send images as `image_url` content parts, interleaved with text if you like. Set `reasoning_effort` to an integer between 1 and 100; low values answer quickly, 100 reasons as far as the model goes. DeepSeek recommends `temperature` 1.0 and `top_p` 0.95. DeepSeek ships its own prompt encoding rather than a chat template; we handle it for you. The Anthropic Messages format is also available. ## DeepSeek, processed in the EU instead of China DeepSeek's own app and API process data in China, which is a hurdle for most European companies. With the open weights that hurdle disappears. On inference.eu, DeepSeek V4.1 Flash runs on infrastructure in the European Union: prompts, images and responses are processed in EU data centres, are not sent to DeepSeek, and are never used to train any model. Every customer gets a data processing agreement under Article 28 of the GDPR and our list of subprocessors, so you can show that inference involves no transfer to a third country. Billing is in euros. The MIT licence allows commercial use with minimal conditions. ## DeepSeek V4.1 Flash (1017) Maker DeepSeek Made in China Type Reasoning Parameters 552B Active parameters 16B Context window 1,048,576 tokens Licence [MIT](https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/LICENSE) Released October 2026 Model ID deepseek-v4.1-flash Processing EU data centres Training on your data Never Source [huggingface.co](https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017) deepseek-v4.1-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="deepseek-v4.1-flash", messages=[{"role": "user", "content": "Summarise this contract in five bullet points."}], ) print(response.choices[0].message.content) ``` ## What teams use it for - 1 ### Agents that read screens Let an agent look at screenshots of web apps or dashboards, decide what to do and call the right tools. - 2 ### Visual document analysis Answer questions about charts, technical drawings and scanned reports alongside long text documents. - 3 ### Cost-tuned reasoning Use a low effort value for routine requests and a high one for hard cases, all with the same model. ## DeepSeek V4.1 Flash (1017): common questions The questions buyers, developers and legal teams ask us most. Something missing? Ask us directly. On inference.eu, DeepSeek V4.1 Flash runs from its openly published weights on EU infrastructure, so no prompts or images reach DeepSeek or China. What remains is the model's behaviour, which reflects its training, including on politically sensitive topics. Evaluate it on your own prompts before production use. Compliance depends on your processing. On inference.eu, 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. The transfer to China that comes with DeepSeek's own service does not happen. DeepSeek-V4.1-Flash-1017 is a minor update released in October 2026. DeepSeek reports better reasoning, coding, agent and visual results, plus shorter outputs and faster task completion. The architecture is the same. On inference.eu, \`deepseek-v4.1-flash\` refers to the 1017 version. DeepSeek V4.1 Flash supports one million tokens of context. Images count towards the context. With high reasoning effort, also leave room for the reasoning trace in the output budget, which can be long at the top of the scale. Yes. It processes images and text natively, using a vision encoder trained together with the language model. You can mix several images and text in one prompt. It generates text only, so it cannot create or edit images. Pass \`reasoning\_effort\` as an integer from 1 to 100. Higher values make the model reason longer and more carefully, at the cost of more output tokens. DeepSeek's published results use 100; for everyday requests a much lower value is usually enough. Yes. The repository and model weights are licensed under the MIT licence, which allows commercial use, modification and redistribution. If you redistribute the weights yourself, keep the copyright and licence notice. V4.1 Flash is newer, multimodal and uses a different architecture with a continuous 1 to 100 effort scale. V4 Flash is text-only with three effort levels and fewer total parameters. Both have a one-million-token context and the MIT licence. Point your OpenAI client's base URL at the inference.eu API, use your inference.eu key and set the model to \`deepseek-v4.1-flash\`. Messages with images, tools and streaming work as before. Add \`reasoning\_effort\` to control the depth of thinking. No. DeepSeek published the weights but is not involved in serving your requests on inference.eu. Your prompts and images are processed in EU data centres to produce your answer and are never used to train DeepSeek or any other model. [Ask a question](https://inference.eu/contact/) ## Related [All models](https://inference.eu/models/#directory) - [DeepSeek V4 FlashRun DeepSeek V4 Flash (0731) on inference.eu: a 1M-token open reasoning model behind an OpenAI-compatible API, processed in EU data centres, not in China.Reasoning](https://inference.eu/models/deepseek-v4-flash/) - [Kimi K3Run Kimi K3, Moonshot AI's 2.8-trillion-parameter open-weight model with 1M context and image and video input, on inference.eu with processing in the EU.Reasoning](https://inference.eu/models/kimi-k3/) - [GLM-5.3-FlashRun Z.ai's GLM-5.3-Flash on inference.eu: a multimodal open model with 18B active parameters and 1M context, OpenAI-compatible, processed in the EU.Reasoning](https://inference.eu/models/glm-5-3-flash/) - [DeepSeek-OCR 2Run DeepSeek-OCR 2 on inference.eu: a 3B open OCR model that turns scans and PDFs into Markdown, processed in EU data centres, not in China.OCR and documents](https://inference.eu/models/deepseek-ocr-2/) ## 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