(https://huggingface.co/front/assets/fonts/source-sans-pro-latin-400-normal.woff2)  (https://huggingface.co/front/assets/fonts/source-sans-pro-latin-600-normal.woff2)  (https://huggingface.co/front/assets/fonts/source-sans-pro-latin-700-normal.woff2)  (../front/build/kube-1f021fd/style.css)  (../front/build/kube-1f021fd/katex.css)  (../front/build/kube-1f021fd/katex.css)   (DeepSeek-V4.1-Flash-1017.html) deepseek-ai/DeepSeek-V4.1-Flash-1017 · Hugging Face (https://huggingface.co/front/build/kube-1f021fd/index.js)      (https://huggingface.co/) (Hugging Face's logo)  Hugging Face  () (Search models, datasets, users...)          (https://huggingface.co/models)     Models   (https://huggingface.co/datasets)      Datasets   (https://huggingface.co/spaces)     Spaces   (https://huggingface.co/storage)     Buckets new   (https://huggingface.co/docs)     Docs   (https://huggingface.co/enterprise)   Enterprise   (https://huggingface.co/pricing) Pricing        Website (https://huggingface.co/tasks)    Tasks   (https://huggingface.co/chat)    HuggingChat   (https://huggingface.co/collections)     Collections   (https://huggingface.co/languages)   Languages   (https://huggingface.co/organizations)      Organizations     Community (https://huggingface.co/blog)   Blog   (https://huggingface.co/posts)   Posts   (https://huggingface.co/papers)    Daily Papers   (https://huggingface.co/hardware)    Hardware   (https://huggingface.co/learn)       Learn   (https://huggingface.co/join/discord)   Discord   (https://discuss.huggingface.co/)   Forum   (https://github.com/huggingface)   GitHub     Solutions (https://huggingface.co/enterprise)   Team & Enterprise   (https://huggingface.co/pro)        Hugging Face PRO   (https://huggingface.co/support)    Enterprise Support   (https://huggingface.co/inference/models)    Inference Providers   (https://huggingface.co/inference-endpoints)    Inference Endpoints   (https://huggingface.co/storage)     Storage Buckets           (https://huggingface.co/login) Log In  (https://huggingface.co/join) Sign Up        (https://huggingface.co/deepseek-ai) ()       (https://huggingface.co/deepseek-ai) deepseek-ai   /  (DeepSeek-V4.1-Flash-1017.html) DeepSeek-V4.1-Flash-1017 (Copy model name to clipboard)       (Like)   Like    (See users who liked this repository) 913    Follow ()  DeepSeek  (Show DeepSeek's followers) 150k    (https://huggingface.co/models?pipeline_tag=image-text-to-text)      Image-Text-to-Text   (https://huggingface.co/models?library=transformers)            Transformers   (https://huggingface.co/models?library=safetensors)   Safetensors   (https://huggingface.co/models?other=deepseek_v41) deepseek_v41   (https://huggingface.co/models?other=text-generation) text-generation   (https://huggingface.co/models?other=8-bit)   8-bit precision   (https://huggingface.co/models?other=fp8) fp8     License: mit       (DeepSeek-V4.1-Flash-1017.html)     Model card (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017/tree/main)    Files Files and versions    xet     (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017/discussions)    Community13               Deploy                 Copy to bucket new      Use this model   Instructions to use deepseek-ai/DeepSeek-V4.1-Flash-1017 with libraries, inference providers, notebooks, and local apps. Follow these links to get started. Libraries (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017?library=transformers)            Transformers  How to use deepseek-ai/DeepSeek-V4.1-Flash-1017 with Transformers: # Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="deepseek-ai/DeepSeek-V4.1-Flash-1017") # pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("deepseek-ai/DeepSeek-V4.1-Flash-1017", device_map="auto")  Notebooks (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017/colab)       Google Colab   (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017/kaggle)    Kaggle   Local Apps (https://huggingface.co/settings/local-apps) (Set up your favorite local applications)   Settings   (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017?local-app=vllm)      vLLM  How to use deepseek-ai/DeepSeek-V4.1-Flash-1017 with vLLM: Install from pip and serve model # Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "deepseek-ai/DeepSeek-V4.1-Flash-1017"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}' Use Docker docker model run hf.co/deepseek-ai/DeepSeek-V4.1-Flash-1017  (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017?local-app=sglang)      SGLang  How to use deepseek-ai/DeepSeek-V4.1-Flash-1017 with SGLang: Install from pip and serve model # Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "deepseek-ai/DeepSeek-V4.1-Flash-1017" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}' Use Docker images docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "deepseek-ai/DeepSeek-V4.1-Flash-1017" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'  (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017?local-app=docker-model-runner)   Docker Model Runner  How to use deepseek-ai/DeepSeek-V4.1-Flash-1017 with Docker Model Runner: docker model run hf.co/deepseek-ai/DeepSeek-V4.1-Flash-1017  (https://huggingface.co/models?other=base_model:quantized:deepseek-ai/DeepSeek-V4.1-Flash-1017)   Browse
							Quantizations to use this model in   llama.cpp ,       Ollama ,           LM Studio , or any compatible app.                   (DeepSeek-V4.1-Flash-1017.html#deepseek-v41-flash-1017) (DeepSeek-V4.1-Flash-1017) DeepSeek-V4.1-Flash-1017 (DeepSeek-V4.1-Flash-1017.html#introduction) (Introduction) Introduction   (DeepSeek-V4.1-Flash-1017.html#prompt-encoding) (Prompt Encoding) Prompt Encoding   (DeepSeek-V4.1-Flash-1017.html#minimal-inference) (Minimal Inference) Minimal Inference   (DeepSeek-V4.1-Flash-1017.html#license) (License) License   (DeepSeek-V4.1-Flash-1017.html#citation) (Citation) Citation   (DeepSeek-V4.1-Flash-1017.html#contact) (Contact) Contact             (DeepSeek-V4.1-Flash-1017.html#deepseek-v41-flash-1017)     DeepSeek-V4.1-Flash-1017   (DeepSeek-V4.1)  (https://www.deepseek.com/) (Homepage)  (https://chat.deepseek.com/) (Chat)   (https://huggingface.co/deepseek-ai) (Hugging Face)  (https://twitter.com/deepseek_ai) (Twitter Follow)   (https://huggingface.co/deepseek-ai/LICENSE) (License)   (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/DeepSeek_V41_Tech_Report.pdf) Technical Report 👁️  (DeepSeek-V4.1-Flash-1017.html#introduction)     Introduction   DeepSeek-V4.1-Flash-1017 is a minor update to (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash) DeepSeek-V4.1-Flash  , delivering substantial improvements across reasoning, coding, agentic tasks, and visual benchmarks. Alongside these gains, the model reduces task execution time and output length, enabling users to complete tasks faster and at lower cost and advancing the performance–cost Pareto frontier across multiple agent harnesses. Benchmark DS-V4.1-Flash-1017 DS-V4.1-Flash K3 GLM-5.3 GPT-5.6 Sol Opus 5   HLE w/o tools 42.6 (44.0†) 36.8 (39.1†) 43.5 42.0† 44.5 56.3  Terminal-Bench 4.0 42.3 31.2 12.6 37.9 39.9 51.8  DeepSWE v1.1 77.0 74.2 67.5 66.9 73.0 74.0  ProgramBench 29.1 20.3 17.5 19.0 23.0 37.0  NL2Repo-Bench 68.2 65.4 58.0 58.0 56.8 75.3  Automation-Bench v1.0.6 60.5 54.8 46.7 48.8 45.8 50.3  Agents’ Last Exam 35.4 31.8 27.6 28.5 26.7 28.6  Apex Bench v1.1 72.1 64.6 47.8 56.6 54.3 65.8  Chartography (w/tools) 83.0 78.9 68.1 — 79.9 84.0  BabyVision (w/tools) 92.5 89.6 85.7 — 88.9 94.1  ZeroBench-main (w/tools) 51.0 49.0 41.0 — 53.0 52.0      † denotes the text-only subset of HLE. We use the Claude Code harness for Terminal-Bench 4.0 and the Mini-SWE-V2 harness for DeepSWE v1.1.  (DeepSeek-V4.1-Flash-1017.html#prompt-encoding)     Prompt Encoding   This release does not include a Jinja-format chat template. The (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017/blob/main/encoding/README.md) encoding  folder contains a self-contained Python reference implementation (encoding.py ) with test cases for multi-turn conversations, tool calling, thinking mode, numeric reasoning effort, mid-conversation system messages, and interleaved image content. For production use, we additionally release (https://github.com/deepseek-ai/deepseek-recipe) deepseek-recipe , a set of Rust libraries with Python bindings that provides the same prompt format as a maintained, protocol-aware toolkit. It converts Messages, Chat Completions, and Responses API requests into the Conversation format, encodes them into DeepSeek V4 and V4.1 prompts or token IDs, and parses model output back into complete or streamed responses — covering thinking, tool calls, images, and generation settings. Model inference, tool execution, and HTTP transport are left to the caller. (DeepSeek-V4.1-Flash-1017.html#minimal-inference)     Minimal Inference   Please refer to the (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017/blob/main/inference/README.md) inference  folder for instructions on weight conversion and running inference locally. Recommended sampling parameters:  Parameter Value   temperature  1.0  top_p  1.0  context_window  1M tokens  max_tokens  ≥ 256K     (DeepSeek-V4.1-Flash-1017.html#license)     License   This repository and the model weights are licensed under the (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017/tree/main/LICENSE) MIT License . (DeepSeek-V4.1-Flash-1017.html#citation)     Citation   @misc{deepseekai2026deepseekv41flash,
      title={DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression},
      author={DeepSeek-AI},
      year={2026},
}   (DeepSeek-V4.1-Flash-1017.html#contact)     Contact   If you have any questions, please raise an issue or contact us at (mailto:service@deepseek.com) service@deepseek.com .    Downloads last month -              Safetensors(https://huggingface.co/docs/safetensors)       Model size 763B params  Tensor type BF16 · F32 · F8_E4M3 · I8 ·     Files info          Inference Providers (https://huggingface.co/docs/inference-providers) NEW    (https://huggingface.co/tasks/image-text-to-text) (Learn more about image-text-to-text)     Image-Text-to-Text      This model isn't deployed by any Inference Provider. (https://huggingface.co/spaces/huggingface/InferenceSupport/discussions/13036) 🙋 13  Ask for provider support            Model tree for deepseek-ai/DeepSeek-V4.1-Flash-1017 (https://huggingface.co/docs/hub/model-cards#specifying-a-base-model)           Finetunes  (https://huggingface.co/models?other=base_model:finetune:deepseek-ai/DeepSeek-V4.1-Flash-1017) 3 models      Quantizations  (https://huggingface.co/models?apps=llama.cpp&other=base_model:quantized:deepseek-ai/DeepSeek-V4.1-Flash-1017) (Use with llama.cpp)    (https://huggingface.co/models?apps=lmstudio&other=base_model:quantized:deepseek-ai/DeepSeek-V4.1-Flash-1017) (Use with LM Studio)            (https://huggingface.co/models?apps=jan&other=base_model:quantized:deepseek-ai/DeepSeek-V4.1-Flash-1017) (Use with Jan)       (https://huggingface.co/models?apps=ollama&other=base_model:quantized:deepseek-ai/DeepSeek-V4.1-Flash-1017) (Use with Ollama)         (https://huggingface.co/models?other=base_model:quantized:deepseek-ai/DeepSeek-V4.1-Flash-1017) 19 models           Spaces using deepseek-ai/DeepSeek-V4.1-Flash-1017 2  (https://huggingface.co/spaces/tardellirs/model-pulse) 📈 tardellirs/model-pulse      (https://huggingface.co/spaces/lalopenguin/deepseek-v4-1-flash-demo) 🐋 lalopenguin/deepseek-v4-1-flash-demo         Collection including deepseek-ai/DeepSeek-V4.1-Flash-1017  (https://huggingface.co/collections/deepseek-ai/deepseek-v4) (DeepSeek-V4) DeepSeek-V4     Collection  11 items • Updated (2026-10-17T05:52:18.716Z) about 17 hours ago  •   932                System theme   Company (https://huggingface.co/terms-of-service) TOS (https://huggingface.co/privacy) Privacy (https://huggingface.co/huggingface) About (https://apply.workable.com/huggingface/) Careers (https://huggingface.co/)                  Website (https://huggingface.co/models) Models (https://huggingface.co/datasets) Datasets (https://huggingface.co/spaces) Spaces (https://huggingface.co/pricing) Pricing (https://huggingface.co/docs) Docs