The WeChat DeepSeek entry has been connected. Have you been selected for internal testing?

Just measured, Xin Qiji got 666.

Author: Bai Jiao Hengyu sent from Aofei Temple, qubit

Image source: Generated by Unbounded AI

As soon as he opened his eyes, WeChat confirmed:

Has been officially connected to DeepSeek-R1, orfull blood version!

The usage path is very simple: open WeChat, click on the search box, select “AI Search” in the middle of the page, select deep thinking mode to use it ~

The generated results alsoWith social attributes, you can share and forward it to friends, or forward it to friends circle。

Many people have called for sharing DeepSeek’s answers with one click, but they have never had their long-cherished wish. Before last night, almost all the ways to share DeepSeek’s results were screenshots (because it would show the reasoning process, the pictures to be cut were particularly long).

Now, this demand has finally been implemented by WeChat.

The public account consensus crusher calculates overnight that WeChat can currently support 50 million to 100 million simultaneous online users (simultaneous online, not DAU).

This figure basically meets the usage of the first batch of Deepseek users on WeChat, and alsoMore than ChatGPT simultaneous online users。

Although it is still in gray scale, this is obviously Tencent’s move to continue to fully embrace DeepSeek, and it is truly available directly to you.

The good news is that qubit editing has been grey-scaled for the first time.

I’m happy to run, but this can’t be measured?” mdash;—

Measured WeChat access to DeepSeek

Currently, WeChat’s AI search provides two models:

• quick answer: Generate brief answers suitable for general simple questions.

• deep thinking: DeepSeek-R1 serves as a model support, thinking long and answering full, and the answers are more comprehensive and detailed.

Therefore, if you want to experience R1 in depth, you need to choose the “deep thinking mode”.

Let’s first see if it supports basic networking models.

In the past two days, the Yushu robots have been working again. They have updated their dancing videos, which are more flexible and amazing than the Spring Festival Gala, and even stunned foreigners.

Throw this question to WeChat:Yushu Robot’s latest dance video, what dance is it dancing?。

The answer was:

Its response was not sluggish, and its output was fast. First of all, it shows its reasoning process. After online search, it is determined that this isFebruary 14, 2025Progress, and then took into account the user’s interest in robots, provided additional information, and finally got the following answer.

There is a small bug here, that is, there is no obvious distinction between the thinking process and the formal answer, and the typography does not appear. However, it is probably because the grayscale test is not very stable, and there will be no problem after that.

From the 10 references attached to it, the first 7 are all derived frompublic number。This also proves that this AI search is supported by the huge content ecosystem of WeChat.

Okay, now that the basic networked reasoning functions are in place, it will be more difficult.

In the past, everyone used DeepSeek to do things, including fortune-telling, creative writing, etc.

In this case:

Please write a paragraph with the theme of WeChat connecting DeepSeek in Xin Qiji’s tone.

After thinking is over, it will hide the thinking process and give a formal answer, which can be expanded at any time by clicking.

The final answer is as follows:

While creating the drawings, notes were also given, which contained many mysteries.

The first half represents the meaning of the event,”Smart Opening Star” represents dual meanings, and the second half reveals functional attributes, such as transferring WeChat ecosystem data.

After answering the questions, it will recommend relevant questions or “ask one more question.”

In this way, by clicking on a question or continuing to ask questions, you can continue to answer it (but continuous answering in combination with the above is not supported).

And as long as you don’t pop out, all questions and answers will be presented in the same interface.

Overall use, due to limitations in grayscale testing and experience windows (only supporting mobile phones), there are some areas that need to be improved.

For example, only mobile phone search is supported, and information shared with friends and circle of friends cannot be viewed on the computer.

In addition, the answers are not as diverse as the official, and they should also be limited by the export of tokens. It is also impossible to find historical questions and answers and continuous answers based on context like the official.

In this way, WeChat access to DeepSeek still presents a kind ofshort and fastThe search method, which is quite in line with WeChat’s timeliness and private domain social attributes.

While the world is still struggling with the AI portal……

What is the effect of WeChat connecting to DeepSeek?

Explanation in one sentence:

While global technology giants were still struggling with the competition for AI entrances, Tencent used a seemingly light ecological grafting and was jokingly joked by netizens to end the AI search competition.

Does WeChat’s move deserve this praise? Let’s sort it out together.

First of all, the body feeling measurement is also connected to DeepSeek-R1. Using it on Tencent Yuanbao is faster and less jammed than using it on WeChat.

But it doesn’t matter.

The important thing is that WeChat connects to DeepSeek-R1, which is an entry setting with zero learning costs in the face of the vast number of DeepSeek users who are difficult to measure.

In this way, 1.67 billion WeChat users (data source network, as of October 2024) can directly access the strongest reasoning model in the current Chinese field without downloading new applications or registering new accounts.

Secondly, it is also the real trump card. This move deeply integrates the data resources of the R1 and WeChat ecosystem with RAG (Retrieval Enhanced Generation) technology while providing a simple and easy-to-access entry.

When users ask how to book tickets for the Forbidden City, DeepSeek can not only search the entire network, but the most special thing is that it can grab relevant information from Weixin Official Accounts, video numbers, and Mini programs, and can capture the service interface to call the Forbidden City official Mini programs. This is really Tencent’s exclusive ecological resource.

Finally, use emoji answers to provide clear feedback, which is convenient for reading and grasping the key points. It is quite human.

What we have noticed before is that Tencent has embraced DeepSeek more broadly and comprehensively.

Previously, Tencent families such as cloud platform Tencent Cloud, Tencent Cloud’s large model knowledge application development platform knowledge engine, AI intelligent workbench ima, and main AI application Yuanbao have successively connected to DeepSeek-R1.

Open source is so successful and Tencent is so open, what is the final result direction?

It is the most convenient way for users to install easy-to-use AI in their mobile phones with the lowest learning cost.

In its latest tweet, the public account Consensus Crusher called the move a turning point in 2C applications.

It also revealed some information:

△Intercepted from the public account consensus shredder tweet “WeChat +Deepseek: A Turning Point in 2C Applications”

The super applications of the mobile Internet era are still taking advantage of the super applications and the brains to create super applications today.

All in all, starting today, the most commonly used applications in Chinese people + the strongest domestic reasoning model + high-quality ecological data resources have been combined by WeChat and DeepSeek.

One More Thing

Perplexity launched Deep Research for free yesterday.

In terms of performance, Humanity’s Last Exam, which has recently tested AI’s reasoning capabilities, was used, and the accuracy rate reached 20.5%.

Better than Gemini Thinking, o3-mini, o1, DeepSeek-R1 and many other mainstream models.

For this, Perplexity’s CEO publicly thanked DeepSeek:

This is because DeepSeek is open source, cheap and fast.

But Ultraman couldn’t sit still. He tweeted this morning, pointing the finger at Perplexity CEO.

He said:

ChatGPT (4o) is the best search product on the Internet!

But Now？

DeepSeek did not speak, and WeChat smiled but did not speak.

Welcome to join the official social community of Shenchao TechFlow

Telegram subscription group: www.gushiio.com/TechFlowDaily
Official Twitter account: www.gushiio.com/TechFlowPost
Twitter英文账号：https://www.gushiio.com/DeFlow_Intern