Total runs:
4.7M
Run Growth:
2.9M
Growth Rate:
61.76%
Sponsored by
APIMart
deepseek-ai / DeepSeek-V4-Pro-0813
Model's Last Updated: August 14 2026
text-generation
Introduction of DeepSeek-V4-Pro-0813
Model Details of DeepSeek-V4-Pro-0813
DeepSeek-V4-Pro-0813
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:
-
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
maxreasoning effort level withtemperature = 1.0, top_p = 0.95. - † 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=?"}
]
# messages -> string
prompt = encode_messages(messages, thinking_mode="thinking", reasoning_effort="max")
# string -> tokens
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 [email protected] .
Runs of deepseek-ai DeepSeek-V4-Pro-0813 on huggingface.co
79.9K
Total runs
-2.6K
24-hour runs
-7.7K
3-day runs
-11.8K
7-day runs
-96.6K
30-day runs
More Information About DeepSeek-V4-Pro-0813 huggingface.co Model
Url of DeepSeek-V4-Pro-0813
DeepSeek-V4-Pro-0813 huggingface.co Url
Provider of DeepSeek-V4-Pro-0813 huggingface.co
Other API from deepseek-ai
Updated:December 01 2025
Total runs:
4.4M
Run Growth:
72.8K
Growth Rate:
1.64%
Total runs:
2.0M
Run Growth:
-331.3K
Growth Rate:
-16.22%
Total runs:
1.5M
Run Growth:
449.8K
Growth Rate:
30.95%
Total runs:
1.3M
Run Growth:
427.8K
Growth Rate:
31.73%
Total runs:
1.3M
Run Growth:
1.2M
Growth Rate:
89.14%
Total runs:
1.3M
Run Growth:
869.0K
Growth Rate:
68.43%
Total runs:
1.1M
Run Growth:
465.9K
Growth Rate:
40.58%
Total runs:
824.9K
Run Growth:
-945.0K
Growth Rate:
-112.12%
Total runs:
769.5K
Run Growth:
7.3K
Growth Rate:
0.92%
Total runs:
688.9K
Run Growth:
-244.7K
Growth Rate:
-35.52%
Total runs:
601.9K
Run Growth:
117.4K
Growth Rate:
19.51%
Total runs:
570.6K
Run Growth:
-388.4K
Growth Rate:
-68.06%
Total runs:
568.2K
Run Growth:
-154.5K
Growth Rate:
-26.57%
Total runs:
399.8K
Run Growth:
-107.8K
Growth Rate:
-26.96%
Total runs:
382.1K
Run Growth:
-237.4K
Growth Rate:
-62.13%
Total runs:
358.7K
Run Growth:
116.6K
Growth Rate:
32.51%
Total runs:
335.4K
Run Growth:
-25.6K
Growth Rate:
-7.63%
Total runs:
330.3K
Run Growth:
-472.1K
Growth Rate:
-142.95%
Total runs:
328.4K
Run Growth:
-16.0K
Growth Rate:
-4.88%
Total runs:
179.5K
Run Growth:
-63.1K
Growth Rate:
-35.13%
Total runs:
169.6K
Run Growth:
-174.1K
Growth Rate:
-102.65%
Total runs:
157.7K
Run Growth:
-154.7K
Growth Rate:
-98.12%
Total runs:
147.5K
Run Growth:
-375.3K
Growth Rate:
-226.94%
Total runs:
113.8K
Run Growth:
-40.3K
Growth Rate:
-35.37%
Total runs:
113.1K
Run Growth:
46.0K
Growth Rate:
40.65%
Total runs:
76.3K
Run Growth:
-17.1K
Growth Rate:
-22.35%
Total runs:
55.0K
Run Growth:
-22.0K
Growth Rate:
-39.98%
Total runs:
51.8K
Run Growth:
-333.2K
Growth Rate:
-515.48%
Total runs:
50.8K
Run Growth:
-8.8K
Growth Rate:
-17.40%
Total runs:
40.3K
Run Growth:
3.9K
Growth Rate:
9.69%
Total runs:
27.8K
Run Growth:
-753
Growth Rate:
-2.71%
Total runs:
27.1K
Run Growth:
-56.4K
Growth Rate:
-207.92%
Total runs:
24.6K
Run Growth:
-3.4K
Growth Rate:
-13.69%
Total runs:
24.5K
Run Growth:
747
Growth Rate:
3.04%
Total runs:
23.2K
Run Growth:
2.3K
Growth Rate:
9.83%
Total runs:
23.1K
Run Growth:
2.5K
Growth Rate:
10.96%
Total runs:
20.6K
Run Growth:
-9.3K
Growth Rate:
-45.05%
Total runs:
20.4K
Run Growth:
-23.1K
Growth Rate:
-113.11%
Total runs:
18.1K
Run Growth:
-4.6K
Growth Rate:
-25.74%
Total runs:
12.9K
Run Growth:
4.9K
Growth Rate:
38.01%
Total runs:
11.3K
Run Growth:
-19.2K
Growth Rate:
-170.06%
Total runs:
11.1K
Run Growth:
1.4K
Growth Rate:
13.01%
Total runs:
10.8K
Run Growth:
2.5K
Growth Rate:
22.99%
Total runs:
10.2K
Run Growth:
-490
Growth Rate:
-4.79%
Total runs:
9.6K
Run Growth:
-4.2K
Growth Rate:
-43.38%
Total runs:
9.4K
Run Growth:
119
Growth Rate:
1.27%
Total runs:
7.7K
Run Growth:
-2.6K
Growth Rate:
-33.49%
Total runs:
7.2K
Run Growth:
-3.0K
Growth Rate:
-40.66%
Total runs:
6.5K
Run Growth:
0
Growth Rate:
0.00%
Total runs:
6.2K
Run Growth:
0
Growth Rate:
0.00%
Total runs:
6.2K
Run Growth:
-1.3K
Growth Rate:
-21.18%
Total runs:
6.1K
Run Growth:
-991
Growth Rate:
-15.94%
Total runs:
5.5K
Run Growth:
-4.2K
Growth Rate:
-76.40%
Total runs:
5.5K
Run Growth:
-3.0K
Growth Rate:
-54.82%
Total runs:
4.9K
Run Growth:
-1.2K
Growth Rate:
-24.98%
Total runs:
4.5K
Run Growth:
1.9K
Growth Rate:
42.25%
Total runs:
4.5K
Run Growth:
0
Growth Rate:
0.00%
Total runs:
3.9K
Run Growth:
-2.5K
Growth Rate:
-63.97%
Total runs:
3.8K
Run Growth:
284
Growth Rate:
7.56%
Total runs:
3.1K
Run Growth:
-17
Growth Rate:
-0.56%
Total runs:
3.0K
Run Growth:
-6.7K
Growth Rate:
-226.54%
Total runs:
2.8K
Run Growth:
-2.7K
Growth Rate:
-88.53%
Total runs:
2.4K
Run Growth:
1.0K
Growth Rate:
42.24%
Total runs:
2.4K
Run Growth:
918
Growth Rate:
37.75%
Total runs:
2.3K
Run Growth:
-499
Growth Rate:
-22.15%
Total runs:
2.1K
Run Growth:
-1.8K
Growth Rate:
-85.60%
Total runs:
1.7K
Run Growth:
-71.7K
Growth Rate:
-4183.32%
Total runs:
1.3K
Run Growth:
-601
Growth Rate:
-47.32%
Total runs:
1.2K
Run Growth:
-753
Growth Rate:
-61.07%
Total runs:
1.1K
Run Growth:
-731
Growth Rate:
-66.70%
Total runs:
1.1K
Run Growth:
110
Growth Rate:
10.23%
Total runs:
937
Run Growth:
320
Growth Rate:
34.15%
Total runs:
924
Run Growth:
148
Growth Rate:
16.02%
Total runs:
834
Run Growth:
0
Growth Rate:
0.00%
Total runs:
755
Run Growth:
-77
Growth Rate:
-10.20%
Total runs:
731
Run Growth:
59
Growth Rate:
8.36%
Total runs:
630
Run Growth:
260
Growth Rate:
41.27%
Total runs:
621
Run Growth:
-7
Growth Rate:
-1.13%
Total runs:
617
Run Growth:
163
Growth Rate:
26.46%
Total runs:
575
Run Growth:
-230
Growth Rate:
-40.00%
Total runs:
566
Run Growth:
5
Growth Rate:
0.88%
Total runs:
561
Run Growth:
-114
Growth Rate:
-20.32%
Total runs:
511
Run Growth:
195
Growth Rate:
38.16%
Total runs:
480
Run Growth:
208
Growth Rate:
43.33%
Total runs:
450
Run Growth:
0
Growth Rate:
0.00%
Total runs:
449
Run Growth:
168
Growth Rate:
37.42%
Total runs:
415
Run Growth:
95
Growth Rate:
22.89%
Total runs:
413
Run Growth:
84
Growth Rate:
20.34%
Total runs:
407
Run Growth:
32
Growth Rate:
7.86%
Total runs:
386
Run Growth:
60
Growth Rate:
15.54%
Total runs:
384
Run Growth:
61
Growth Rate:
15.89%
Total runs:
384
Run Growth:
50
Growth Rate:
13.70%
Total runs:
375
Run Growth:
60
Growth Rate:
16.00%
Total runs:
358
Run Growth:
35
Growth Rate:
9.78%
Total runs:
337
Run Growth:
5
Growth Rate:
1.62%
Total runs:
332
Run Growth:
0
Growth Rate:
0.00%
Total runs:
326
Run Growth:
-5
Growth Rate:
-1.67%
Total runs:
311
Run Growth:
-15
Growth Rate:
-5.23%