AI News
开源项目 Hugging Face Models 2026-08-14T00:01 正文 ok 653 words

DeepSeek-V4-Pro-0813：正式版发布，Agent 能力与生产环境性能显著增强

英文标题：🤗 deepseek-ai/DeepSeek-V4-Pro-0813

DeepSeek 发布 DeepSeek-V4-Pro-0813，作为 DeepSeek-V4-Pro 预览版的正式替代版本，在既有模型结构上集成了 DSpark 推测解码模块。官方称，该模型在多项代码 Agent、工具使用和自动化基准上优于预览版，并与当前最强的闭源模型展开竞争。面向开发者，模型新增 `low`、`high` 和 `max` 三档 `reasoning_effort`，并提供 OpenAI 兼容消息编码脚本以及 vLLM、SGLang 和本地部署指南。对于高推理强度任务，官方建议将最大输出长度设置为 384K tokens。
打开原文相关 1.6 · 机会 2
media4.5 KB media3.6 KB media10.4 KB media6.0 KB media1.3 KB media1.3 KB media1.8 KB media1.3 KB media412.7 KB

中文

DeepSeek-V4-Pro-0813

技术报告 👁️

简介

DeepSeek-V4-Pro-0813 是 DeepSeek-V4-Pro 的正式发布版本，取代此前的预览版本，Agent 能力获得大幅增强，性能也有所提升，且在生产环境中的提升尤为明显。该模型基于 DeepSeek-V4-Pro（Preview）的模型结构构建，并附加了 DSpark 推测解码模块。

DeepSeek-V4-Pro-0813 在下列基准测试中优于 DeepSeek-V4-Pro（Preview），并且整体上能够与当前最强的闭源模型竞争。

基准测试 DeepSeek-V4-Pro-0813 DeepSeek-V4-Flash-0731 DeepSeek-V4-Pro（Preview） DeepSeek-V4-Flash（Preview） GLM-5.2 Kimi K3 Opus-4.8 Fable-5（含回退）
HLE（无工具 / 使用工具） 42.7 / 60.0 37.8 / 51.5 37.7 / 48.2 34.8 / 45.1 40.5 / 54.7 43.5 / 56.0 49.8 / 57.9 53.3 / 63.0
Terminal Bench 2.1 87.9 82.7 72.1 61.8 81.0 88.3 85.0 88.0
NL2Repo 61.5 54.2 38.5 39.4 48.9 - 69.7 -
Cybergym 83.3 76.7 52.7 38.7 - 80.0 78.3 83.1
DeepSWE 62.7 54.4 12.8 7.3 46.2 67.5 58.0 70.0
Toolathlon-Verified 74.1 70.3 55.9 49.7 59.9 76.5 76.2 77.9
Agents' Last Exam 25.7 25.2 16.5 15.8 23.8 27.6 25.7 -
AutomationBench（公开） 31.8 25.1 12.8 10.8 12.9 30.8 27.2 29.1
DSBench-FullStack † 71.1 68.7 41.8 37.0 61.8 73.7 71.6 77.2
DSBench-Hard † 67.2 59.6 31.1 25.8 54.5 63.0 71.7 68.3

注：

  1. 在上述公开基准中的代码 Agent 任务上，DeepSeek-V4-Pro-0813 使用 DeepSeek Harness 的最简模式作为 Agent 框架进行评测，并采用 max 推理强度，参数设置为 temperature = 1.0, top_p = 0.95。
  2. † DSBench-FullStack 是一个内部全栈开发测试集；DSBench-Hard 是一个内部高难度代码 Agent 问题测试集。

聊天模板

本次发布不包含 Jinja 格式的聊天模板。我们改为提供专用的 encoding 文件夹，其中包含 Python 脚本和测试用例，用于展示如何将 OpenAI 兼容格式的消息编码为模型输入字符串，以及如何解析模型的文本输出。完整文档请参阅 encoding 文件夹。

reasoning_effort 参数现支持三个等级：low、high 和 max，用于控制模型在回答前进行思考的程度。

简要示例：

from encoding_dsv4 import encode_messages, parse_message_from_completion_text

messages = [
    {"role": "user", "content": "hello"},
    {"role": "assistant", "content": "Hello! I am DeepSeek.", "reasoning_content": "thinking..."},
    {"role": "user", "content": "1+1=?"}
]


prompt = encode_messages(messages, thinking_mode="thinking", reasoning_effort="max")


import transformers
tokenizer = transformers.AutoTokenizer.from_pretrained("deepseek-ai/DeepSeek-V4-Pro-0813")
tokens = tokenizer.encode(prompt)

如何使用 vLLM 运行

通过一个标志即可启用 DSpark 推测解码——在 vLLM 启动命令中添加带有 method: dspark 的 --speculative-config：

--speculative-config '{"method":"dspark","num_speculative_tokens":7,"draft_sample_method":"greedy"}'

例如，以下命令可在单个 4×GB300 节点上通过 vLLM 提供该模型服务。有关详细说明及其他硬件配置，请参阅 vLLM 配方。

vllm serve deepseek-ai/DeepSeek-V4-Pro-0813 \
  --trust-remote-code --kv-cache-dtype fp8 --block-size 256 \
  --data-parallel-size 4 --enable-expert-parallel \
  --moe-backend deep_gemm_mega_moe \
  --attention-config '{"use_fp4_indexer_cache": true}' \
  --speculative-config '{"method":"dspark","num_speculative_tokens":7,"draft_sample_method":"greedy"}'

如何使用 SGLang 运行

使用 --speculative-algorithm DSPARK 启用 DSpark，且不要单独设置 --speculative-draft-model-path，因为目标模型与草稿模型的权重来自同一个检查点。有关详细说明、基准测试和其他硬件配置，请参阅 SGLang cookbook。

sglang serve \
  --trust-remote-code \
  --model-path deepseek-ai/DeepSeek-V4-Pro-0813 \
  --tp 4 \
  --moe-runner-backend flashinfer_mxfp4 \
  --speculative-algorithm DSPARK \
  --mem-fraction-static 0.90 \
  --chunked-prefill-size 4096 \
  --swa-full-tokens-ratio 0.1 \

如何本地运行

有关在本地运行 DeepSeek-V4 的详细说明，包括模型权重转换和交互式聊天演示，请参阅 inference 文件夹。

对于本地部署，我们建议将采样参数设置为 temperature = 1.0；在 Agent 场景中使用 top_p = 0.95，其他场景则使用 top_p = 1.0。对于 high 和 max 推理强度等级，建议将最大输出长度设为 384K tokens。

许可证

本仓库及模型权重采用 MIT License 授权。

引用

@misc{deepseekai2026deepseekv4,
      title={DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence},
      author={DeepSeek-AI},
      year={2026},
}

联系方式

如有任何问题，请提交 issue，或通过 service@deepseek.com 联系我们。

English / 原文

DeepSeek-V4-Pro-0813

Technical Report 👁️

Introduction

DeepSeek-V4-Pro-0813 is the official release of DeepSeek-V4-Pro , superseding the preview version, with greatly enhanced agentic capabilities and performance improvements that are especially pronounced in production environments. It is built on the DeepSeek-V4-Pro (Preview) model structure, with a DSpark speculative decoding module attached.

DeepSeek-V4-Pro-0813 outperforms DeepSeek-V4-Pro (Preview) on the benchmarks listed below, and is broadly competitive with the strongest proprietary models available.

Benchmark DeepSeek-V4-Pro-0813 DeepSeek-V4-Flash-0731 DeepSeek-V4-Pro (Preview) DeepSeek-V4-Flash (Preview) GLM-5.2 Kimi K3 Opus-4.8 Fable-5 (w/ fallback)
HLE (wo / w tools) 42.7 / 60.0 37.8 / 51.5 37.7 / 48.2 34.8 / 45.1 40.5 / 54.7 43.5 / 56.0 49.8 / 57.9 53.3 / 63.0
Terminal Bench 2.1 87.9 82.7 72.1 61.8 81.0 88.3 85.0 88.0
NL2Repo 61.5 54.2 38.5 39.4 48.9 - 69.7 -
Cybergym 83.3 76.7 52.7 38.7 - 80.0 78.3 83.1
DeepSWE 62.7 54.4 12.8 7.3 46.2 67.5 58.0 70.0
Toolathlon-Verified 74.1 70.3 55.9 49.7 59.9 76.5 76.2 77.9
Agents' Last Exam 25.7 25.2 16.5 15.8 23.8 27.6 25.7 -
AutomationBench (Public) 31.8 25.1 12.8 10.8 12.9 30.8 27.2 29.1
DSBench-FullStack † 71.1 68.7 41.8 37.0 61.8 73.7 71.6 77.2
DSBench-Hard † 67.2 59.6 31.1 25.8 54.5 63.0 71.7 68.3

Notes:

  1. For the code-agent tasks among the public benchmarks above, DeepSeek-V4-Pro-0813 is evaluated with the minimal mode of DeepSeek Harness as the agent framework, using the max reasoning effort level with temperature = 1.0, top_p = 0.95.
  2. † DSBench-FullStack is an internal full-stack development test set; DSBench-Hard is an internal test set of difficult coding-agent problems.

Chat Template

This release does not include a Jinja-format chat template. Instead, we provide a dedicated encoding folder with Python scripts and test cases demonstrating how to encode messages in OpenAI-compatible format into input strings for the model, and how to parse the model's text output. Please refer to the encoding folder for full documentation.

The reasoning_effort parameter now supports three levels — low, high, and max — which control how much deliberation the model spends before answering.

A brief example:

from encoding_dsv4 import encode_messages, parse_message_from_completion_text

messages = [
    {"role": "user", "content": "hello"},
    {"role": "assistant", "content": "Hello! I am DeepSeek.", "reasoning_content": "thinking..."},
    {"role": "user", "content": "1+1=?"}
]


prompt = encode_messages(messages, thinking_mode="thinking", reasoning_effort="max")


import transformers
tokenizer = transformers.AutoTokenizer.from_pretrained("deepseek-ai/DeepSeek-V4-Pro-0813")
tokens = tokenizer.encode(prompt)

How to Run with vLLM

DSpark speculative decoding is enabled with a single flag — add --speculative-config with method: dspark to your vLLM launch command:

--speculative-config '{"method":"dspark","num_speculative_tokens":7,"draft_sample_method":"greedy"}'

For example, the command below serves the model with vLLM on a single 4×GB300 node. See the vLLM recipe for detailed instructions and other hardware configurations.

vllm serve deepseek-ai/DeepSeek-V4-Pro-0813 \
  --trust-remote-code --kv-cache-dtype fp8 --block-size 256 \
  --data-parallel-size 4 --enable-expert-parallel \
  --moe-backend deep_gemm_mega_moe \
  --attention-config '{"use_fp4_indexer_cache": true}' \
  --speculative-config '{"method":"dspark","num_speculative_tokens":7,"draft_sample_method":"greedy"}'

How to Run with SGLang

Enable DSpark with --speculative-algorithm DSPARK and do not set a separate --speculative-draft-model-path as the target and draft weights therefore come from the same checkpoint. See the SGLang cookbook for detailed instructions, benchmarks and other hardwares configurations.

sglang serve \
  --trust-remote-code \
  --model-path deepseek-ai/DeepSeek-V4-Pro-0813 \
  --tp 4 \
  --moe-runner-backend flashinfer_mxfp4 \
  --speculative-algorithm DSPARK \
  --mem-fraction-static 0.90 \
  --chunked-prefill-size 4096 \
  --swa-full-tokens-ratio 0.1 \

How to Run Locally

Please refer to the inference folder for detailed instructions on running DeepSeek-V4 locally, including model weight conversion and interactive chat demos.

For local deployment, we recommend setting the sampling parameters to temperature = 1.0, with top_p = 0.95 for agentic scenarios and top_p = 1.0 otherwise. For the high and max reasoning effort levels, we recommend a maximum output length of 384K tokens.

License

This repository and the model weights are licensed under the MIT License.

Citation

@misc{deepseekai2026deepseekv4,
      title={DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence},
      author={DeepSeek-AI},
      year={2026},
}

Contact

If you have any questions, please raise an issue or contact us at service@deepseek.com.