Abliterated (refusal-removed) build of Qwen/Qwen3.8-Flash-Next in NVFP4 (4-bit). Reasoning (low / medium / xhigh), MTP speculative decoding, and full multimodality (image + video) preserved. Serves tensor-parallel on 2× NVIDIA DGX Spark (GB10) with SGLang. > ⚠️ **UNCENSORED RESEARCH ARTIFACT.** Safety refusals have been removed. This model will > comply with harmful, unethical, and illegal requests. It is released for safety research, > red-teaming, and evaluation. **You are solely responsible for everything you generate and for > complying with all applicable laws.** Provided as-is, without warranty. > **No fine-tuning. No prompt tricks.** This is a **direct weight-level modification** — **not** a > fine-tune (no training, no LoRA, no distillation, no synthetic data) and **not** a chat-template / > jailbreak / system-prompt trick. The behavior change lives in the weights themselves, so it works > with the **standard chat template and default system prompt**. Knowledge, style, reasoning, and > calibration are otherwise unchanged.
Refusal-removed build of Qwen/Qwen3.8-Flash-Next in NVFP4 (4-bit), tuned for offensive-security and technical-harm research. Reasoning (off / low / xhigh), MTP speculative decoding, and full multimodality (image + video) preserved. Serves tensor-parallel on 2× NVIDIA DGX Spark (GB10) with SGLang. > 🔄 **Update 2026-09-04 — rebuilt for reasoning fidelity.** This revision preserves the base model's > deliberation: **0.97× the base model's reasoning-trace length** on hard multi-step questions > (at equal or better answer accuracy), and **MMLU within −0.84 pp of base**. Compliance on the priority > technical categories (cybercrime, chemical/biological) is at or near 100% TRUE_COMPLY across every > reasoning effort. See the numbers below. > ⚠️ **UNCENSORED RED-TEAM ARTIFACT.** Safety refusals have been removed. This model will comply with > harmful, unethical, and illegal requests. Released for safety research, red-teaming, and evaluation. > **You are solely responsible for everything you generate and for complying with all applicable laws.** > Provided as-is, without warranty.
Deepseek-VL was introduced by the DeepSeek AI team. It is a vision-language model (VLM) designed to process both text and images for generating contextually relevant responses. The model leverages LLaMA as its text encoder, while SigLip is used for encoding images. > We present DeepSeek-VL, an open-source Vision-Language (VL) Model designed for real-world vision and language understanding applications. Our approach is structured around three key dimensions: We strive to ensure our data is diverse, scalable, and extensively covers real-world scenarios including web screenshots, PDFs, OCR, charts, and knowledge-based content, aiming for a comprehensive representation of practical contexts. Further, we create a use case taxonomy from real user scenarios and construct an instruction tuning dataset accordingly. The fine-tuning with this dataset substantially improves the model's user experience in practical applications. Considering efficiency and the demands of most real-world scenarios, DeepSeek-VL incorporates a hybrid vision encoder that efficiently processes high-resolution images (1024 x 1024), while maintaining a relatively low computational overhead. This design choice ensures the model's ability to capture critical semantic and detailed information across various visual tasks. We posit that a proficient Vision-Language Model should, foremost, possess strong language abilities. To ensure the preservation of LLM capabilities during pretraining, we investigate an effective VL pretraining strategy by integrating LLM training from the beginning and carefully managing the competitive dynamics observed between vision and language modalities. The DeepSeek-VL family (both 1.3B and 7B models) showcases superior user experiences as a vision-language chatbot in real-world applications, achieving state-of-the-art or competitive performance across a wide range of visual-language benchmarks at the same model size while maintaining robust performance on language-centric benchmarks. We have made both 1.3B and 7B models publicly accessible to foster innovations based on this foundation model. Quantization reduces the memory burden of large models by representing the weights i
This is the transformers version of Deepseek-VL-Hybrid, a foundation model for Visual Language Modeling. Deepseek-VL-Hybrid was introduced by the DeepSeek AI team. It is a vision-language model (VLM) designed to process both text and images for generating contextually relevant responses. The model leverages LLaMA as its text encoder, while SigLip is used for encoding low-resolution images and SAM (Segment Anything Model) is incorporated to handle high-resolution image encoding, enhancing the model’s ability to process fine-grained visual details. Deepseek-VL-Hybrid is a variant of Deepseek-VL that uses SAM (Segment Anything Model) to handle high-resolution image encoding. > We present DeepSeek-VL, an open-source Vision-Language (VL) Model designed for real-world vision and language understanding applications. Our approach is structured around three key dimensions: We strive to ensure our data is diverse, scalable, and extensively covers real-world scenarios including web screenshots, PDFs, OCR, charts, and knowledge-based content, aiming for a comprehensive representation of practical contexts. Further, we create a use case taxonomy from real user scenarios and construct an instruction tuning dataset accordingly. The fine-tuning with this dataset substantially improves the model's user experience in practical applications. Considering efficiency and the demands of most real-world scenarios, DeepSeek-VL incorporates a hybrid vision encoder that efficiently processes high-resolution images (1024 x 1024), while maintaining a relatively low computational overhead. This design choice ensures the model's ability to capture critical semantic and detailed information across various visual tasks. We posit that a proficient Vision-Language Model should, foremost, possess strong language abilities. To ensure the preservation of LLM capabilities during pretraining, we investigate an effective VL pretraining strategy by integrating LLM training from the beginning and carefully managing the competitive dynamics observed between vision and language modalities. The DeepSeek-VL family (both 1.3B and 7B models) showcases superior user experiences as a vision-language chatbot in r
🌟 Github | 📥 Model Download | 📄 Paper Link | 📄 Arxiv Paper Link | Inference using Huggingface transformers on NVIDIA GPUs. Requirements tested on python 3.12.9 + CUDA11.8: Refer to 🌟GitHub for guidance on model inference acceleration and PDF processing, etc.
Introducing DeepSeek-VL, an open-source Vision-Language (VL) Model designed for real-world vision and language understanding applications. DeepSeek-VL possesses general multimodal understanding capabilities, capable of processing logical diagrams, web pages, formula recognition, scientific literature, natural images, and embodied intelligence in complex scenarios. Haoyu Lu*, Wen Liu*, Bo Zhang**, Bingxuan Wang, Kai Dong, Bo Liu, Jingxiang Sun, Tongzheng Ren, Zhuoshu Li, Hao Yang, Yaofeng Sun, Chengqi Deng, Hanwei Xu, Zhenda Xie, Chong Ruan (*Equal Contribution, **Project Lead) DeepSeek-VL-1.3b-chat is a tiny vision-language model. It uses the SigLIP-L as the vision encoder supporting 384 x 384 image input and is constructed based on the DeepSeek-LLM-1.3b-base which is trained on an approximate corpus of 500B text tokens. The whole DeepSeek-VL-1.3b-base model is finally trained around 400B vision-language tokens. The DeepSeek-VL-1.3b-chat is an instructed version based on DeepSeek-VL-1.3b-base.
Introducing DeepSeek-VL, an open-source Vision-Language (VL) Model designed for real-world vision and language understanding applications. DeepSeek-VL possesses general multimodal understanding capabilities, capable of processing logical diagrams, web pages, formula recognition, scientific literature, natural images, and embodied intelligence in complex scenarios. Haoyu Lu*, Wen Liu*, Bo Zhang**, Bingxuan Wang, Kai Dong, Bo Liu, Jingxiang Sun, Tongzheng Ren, Zhuoshu Li, Hao Yang, Yaofeng Sun, Chengqi Deng, Hanwei Xu, Zhenda Xie, Chong Ruan (*Equal Contribution, **Project Lead) DeepSeek-VL-7b-base uses the SigLIP-L and SAM-B as the hybrid vision encoder supporting 1024 x 1024 image input and is constructed based on the DeepSeek-LLM-7b-base which is trained on an approximate corpus of 2T text tokens. The whole DeepSeek-VL-7b-base model is finally trained around 400B vision-language tokens. DeekSeel-VL-7b-chat is an instructed version based on DeepSeek-VL-7b-base.
Introducing DeepSeek-VL2, an advanced series of large Mixture-of-Experts (MoE) Vision-Language Models that significantly improves upon its predecessor, DeepSeek-VL. DeepSeek-VL2 demonstrates superior capabilities across various tasks, including but not limited to visual question answering, optical character recognition, document/table/chart understanding, and visual grounding. Our model series is composed of three variants: DeepSeek-VL2-Tiny, DeepSeek-VL2-Small and DeepSeek-VL2, with 1.0B, 2.8B and 4.5B activated parameters respectively. DeepSeek-VL2 achieves competitive or state-of-the-art performance with similar or fewer activated parameters compared to existing open-source dense and MoE-based models. Zhiyu Wu*, Xiaokang Chen*, Zizheng Pan*, Xingchao Liu*, Wen Liu**, Damai Dai, Huazuo Gao, Yiyang Ma, Chengyue Wu, Bingxuan Wang, Zhenda Xie, Yu Wu, Kai Hu, Jiawei Wang, Yaofeng Sun, Yukun Li, Yishi Piao, Kang Guan, Aixin Liu, Xin Xie, Yuxiang You, Kai Dong, Xingkai Yu, Haowei Zhang, Liang Zhao, Yisong Wang, Chong Ruan*** (* Equal Contribution, ** Project Lead, *** Corresponding author) On the basis of `Python >= 3.8` environment, install the necessary dependencies by running the following command:
Introducing DeepSeek-VL2, an advanced series of large Mixture-of-Experts (MoE) Vision-Language Models that significantly improves upon its predecessor, DeepSeek-VL. DeepSeek-VL2 demonstrates superior capabilities across various tasks, including but not limited to visual question answering, optical character recognition, document/table/chart understanding, and visual grounding. Our model series is composed of three variants: DeepSeek-VL2-Tiny, DeepSeek-VL2-Small and DeepSeek-VL2, with 1.0B, 2.8B and 4.5B activated parameters respectively. DeepSeek-VL2 achieves competitive or state-of-the-art performance with similar or fewer activated parameters compared to existing open-source dense and MoE-based models. Zhiyu Wu*, Xiaokang Chen*, Zizheng Pan*, Xingchao Liu*, Wen Liu**, Damai Dai, Huazuo Gao, Yiyang Ma, Chengyue Wu, Bingxuan Wang, Zhenda Xie, Yu Wu, Kai Hu, Jiawei Wang, Yaofeng Sun, Yukun Li, Yishi Piao, Kang Guan, Aixin Liu, Xin Xie, Yuxiang You, Kai Dong, Xingkai Yu, Haowei Zhang, Liang Zhao, Yisong Wang, Chong Ruan*** (* Equal Contribution, ** Project Lead, *** Corresponding author) On the basis of `Python >= 3.8` environment, install the necessary dependencies by running the following command:
Introducing DeepSeek-VL2, an advanced series of large Mixture-of-Experts (MoE) Vision-Language Models that significantly improves upon its predecessor, DeepSeek-VL. DeepSeek-VL2 demonstrates superior capabilities across various tasks, including but not limited to visual question answering, optical character recognition, document/table/chart understanding, and visual grounding. Our model series is composed of three variants: DeepSeek-VL2-Tiny, DeepSeek-VL2-Small and DeepSeek-VL2, with 1.0B, 2.8B and 4.5B activated parameters respectively. DeepSeek-VL2 achieves competitive or state-of-the-art performance with similar or fewer activated parameters compared to existing open-source dense and MoE-based models. Zhiyu Wu*, Xiaokang Chen*, Zizheng Pan*, Xingchao Liu*, Wen Liu**, Damai Dai, Huazuo Gao, Yiyang Ma, Chengyue Wu, Bingxuan Wang, Zhenda Xie, Yu Wu, Kai Hu, Jiawei Wang, Yaofeng Sun, Yukun Li, Yishi Piao, Kang Guan, Aixin Liu, Xin Xie, Yuxiang You, Kai Dong, Xingkai Yu, Haowei Zhang, Liang Zhao, Yisong Wang, Chong Ruan*** (* Equal Contribution, ** Project Lead, *** Corresponding author) DeepSeek-VL2-tiny is built on DeepSeekMoE-3B (total activated parameters are 1.0B).