---
title: "Discussion on \"\"Three Chinese LLMS Overseas Latency Tests: DeepSeek 1.45 Seconds is the Fastest. Do You Have a better Testing Method?\"\" | Hashnode"
description: "Discussion on \"\"Three Chinese LLMS Overseas Latency Tests: DeepSeek 1.45 Seconds is the Fastest. Do You Have a better Testing Method?\"\". A Chinese teacher's real test record of transforming into an API entrepreneur | Free methods disclosed, sincerely inviting tech experts to offer guidance"
url: "https://hashnode.com/posts/three-chinese-llms-overseas-latency-tests-deepseek-1-45-seconds-is-the-fastest-do-you-have-a-better-testing-method/69ef1cf1dc472baafda37194"
---

[Feed](https://hashnode.com/feed) /

Discussion

# ["Three Chinese LLMS Overseas Latency Tests: DeepSeek 1.45 Seconds is the Fastest. Do You Have a better Testing Method?"](https://xiaoqiangapi.hashnode.dev/three-chinese-llms-overseas-latency-tests-deepseek-1-45-seconds-is-the-fastest-do-you-have-a-better-testing-method?utm_source=hashnode&utm_medium=feed)

A Chinese teacher's real test record of transforming into an API entrepreneur | Free methods disclosed, sincerely inviting tech experts to offer guidance

Apr 27 · 8 min read · xiaoqiangapi.hashnode.dev  
Updated Apr 30  
Read the article

-   [#api](https://hashnode.com/tag/api)
-   [#deepseek](https://hashnode.com/tag/deepseek)
-   [#latency](https://hashnode.com/tag/latency)
-   [#buildinpublic](https://hashnode.com/tag/buildinpublic)

## Written by

[xiaoqiangapi](https://hashnode.com/@xiaoqiangapi)

@xiaoqiangapi · Singapore

  
[View profile](https://hashnode.com/@xiaoqiangapi)

Chinese teacher → indie maker. Running 3 LLMs. Sharing my build in public journey.

Chinese teacher → indie maker. Providing AI API access (DeepSeek, Zhipu, MiniMax) , Learning in public. #buildinpublic

## Responses 1

Markdown supported · ⌘↵ to post

[xiaoqiangapi](https://hashnode.com/@xiaoqiangapi)

[Apr 27](https://hashnode.com/posts/three-chinese-llms-overseas-latency-tests-deepseek-1-45-seconds-is-the-fastest-do-you-have-a-better-testing-method/69ef1cf1dc472baafda37194/comment/69ef20c7dc472baafda66d1b)

Thank you for reading! This article is a true record of my test as a liberal arts student making a transition. The methods might be a bit "clumsy", but all the data are honest.

If you have more professional API latency testing methods (such as using scripts to measure TTFT, conducting concurrent pressure tests, etc.), or if you have encountered similar speed issues when using Chinese LLMs, we would be very glad to have you share your experience in the comment section.

I will also continue to update more comprehensive comparison data based on everyone's feedback. Thank you for your support 🙏
