DeepSeek-V3 685B

1. Introduction

We present DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi-head Latent Attention (MLA) and DeepSeekMoE architectures, which were thoroughly validated in DeepSeek-V2. Furthermore, DeepSeek-V3 pioneers an auxiliary-loss-free strategy for load balancing and sets a multi-token prediction training objective for stronger performance. We pre-train DeepSeek-V3 on 14.8 trillion diverse and high-quality tokens, followed by Supervised Fine-Tuning and Reinforcement Learning stages to fully harness its capabilities. Comprehensive evaluations reveal that DeepSeek-V3 outperforms other open-source models and achieves performance comparable to leading closed-source models. Despite its excellent performance, DeepSeek-V3 requires only 2.788M H800 GPU hours for its full training. In addition, its training process is remarkably stable. Throughout the entire training process, we did not experience any irrecoverable loss spikes or perform any rollbacks.

2. Model Summary

Architecture: Innovative Load Balancing Strategy and Training Objective

On top of the efficient architecture of DeepSeek-V2, we pioneer an auxiliary-loss-free strategy for load balancing, which minimizes the performance degradation that arises from encouraging load balancing.
We investigate a Multi-Token Prediction (MTP) objective and prove it beneficial to model performance. It can also be used for speculative decoding for inference acceleration.

Pre-Training: Towards Ultimate Training Efficiency

We design an FP8 mixed precision training framework and, for the first time, validate the feasibility and effectiveness of FP8 training on an extremely large-scale model.
Through co-design of algorithms, frameworks, and hardware, we overcome the communication bottleneck in cross-node MoE training, nearly achieving full computation-communication overlap.
This significantly enhances our training efficiency and reduces the training costs, enabling us to further scale up the model size without additional overhead.
At an economical cost of only 2.664M H800 GPU hours, we complete the pre-training of DeepSeek-V3 on 14.8T tokens, producing the currently strongest open-source base model. The subsequent training stages after pre-training require only 0.1M GPU hours.

Post-Training: Knowledge Distillation from DeepSeek-R1

We introduce an innovative methodology to distill reasoning capabilities from the long-Chain-of-Thought (CoT) model, specifically from one of the DeepSeek R1 series models, into standard LLMs, particularly DeepSeek-V3. Our pipeline elegantly incorporates the verification and reflection patterns of R1 into DeepSeek-V3 and notably improves its reasoning performance. Meanwhile, we also maintain a control over the output style and length of DeepSeek-V3.

3. Model Downloads

NOTE: The total size of DeepSeek-V3 models on HuggingFace is 685B, which includes 671B of the Main Model weights and 14B of the Multi-Token Prediction (MTP) Module weights.

To ensure optimal performance and flexibility, we have partnered with open-source communities and hardware vendors to provide multiple ways to run the model locally. For step-by-step guidance, check out Section 6: How_to Run_Locally.

For developers looking to dive deeper, we recommend exploring README_WEIGHTS.md for details on the Main Model weights and the Multi-Token Prediction (MTP) Modules. Please note that MTP support is currently under active development within the community, and we welcome your contributions and feedback.

4. Evaluation Results

Base Model

Standard Benchmarks

Note: Best results are shown in bold. Scores with a gap not exceeding 0.3 are considered to be at the same level. DeepSeek-V3 achieves the best performance on most benchmarks, especially on math and code tasks. For more evaluation details, please check our paper.

Context Window

Evaluation results on the Needle In A Haystack (NIAH) tests. DeepSeek-V3 performs well across all context window lengths up to 128K.

Chat Model

Standard Benchmarks (Models larger than 67B)

Open Ended Generation Evaluation

5. Chat Website & API Platform

You can chat with DeepSeek-V3 on DeepSeek's official website: chat.deepseek.com

We also provide OpenAI-Compatible API at DeepSeek Platform: platform.deepseek.com

6. How to Run Locally

DeepSeek-V3 can be deployed locally using the following hardware and open-source community software:

DeepSeek-Infer Demo: We provide a simple and lightweight demo for FP8 and BF16 inference.
SGLang: Fully support the DeepSeek-V3 model in both BF16 and FP8 inference modes.
LMDeploy: Enables efficient FP8 and BF16 inference for local and cloud deployment.
TensorRT-LLM: Currently supports BF16 inference and INT4/8 quantization, with FP8 support coming soon.
vLLM: Support DeekSeek-V3 model with FP8 and BF16 modes for tensor parallelism and pipeline parallelism.
AMD GPU: Enables running the DeepSeek-V3 model on AMD GPUs via SGLang in both BF16 and FP8 modes.
Huawei Ascend NPU: Supports running DeepSeek-V3 on Huawei Ascend devices.

Since FP8 training is natively adopted in our framework, we only provide FP8 weights. If you require BF16 weights for experimentation, you can use the provided conversion script to perform the transformation.

Here is an example of converting FP8 weights to BF16:

cd inference
python fp8_cast_bf16.py --input-fp8-hf-path /path/to/fp8_weights --output-bf16-hf-path /path/to/bf16_weights

NOTE: Huggingface's Transformers has not been directly supported yet.

6.1 Inference with DeepSeek-Infer Demo (example only)

Model Weights & Demo Code Preparation

First, clone our DeepSeek-V3 GitHub repository:

git clone https://github.com/deepseek-ai/DeepSeek-V3.git

Navigate to the inference folder and install dependencies listed in requirements.txt.

cd DeepSeek-V3/inference
pip install -r requirements.txt

Download the model weights from HuggingFace, and put them into /path/to/DeepSeek-V3 folder.

Model Weights Conversion

Convert HuggingFace model weights to a specific format:

python convert.py --hf-ckpt-path /path/to/DeepSeek-V3 --save-path /path/to/DeepSeek-V3-Demo --n-experts 256 --model-parallel 16

Run

Then you can chat with DeepSeek-V3:

torchrun --nnodes 2 --nproc-per-node 8 generate.py --node-rank $RANK --master-addr $ADDR --ckpt-path /path/to/DeepSeek-V3-Demo --config configs/config_671B.json --interactive --temperature 0.7 --max-new-tokens 200

Or batch inference on a given file:

torchrun --nnodes 2 --nproc-per-node 8 generate.py --node-rank $RANK --master-addr $ADDR --ckpt-path /path/to/DeepSeek-V3-Demo --config configs/config_671B.json --input-file $FILE

6.2 Inference with SGLang (recommended)

SGLang currently supports MLA optimizations, FP8 (W8A8), FP8 KV Cache, and Torch Compile, delivering state-of-the-art latency and throughput performance among open-source frameworks.

Notably, SGLang v0.4.1 fully supports running DeepSeek-V3 on both NVIDIA and AMD GPUs, making it a highly versatile and robust solution.

Here are the launch instructions from the SGLang team: https://github.com/sgl-project/sglang/tree/main/benchmark/deepseek_v3

...

Model Card