After testing the DeepSeek R1 officially supported by WeChat, I figured out how Tencent plans to bet on AI applications

Since its launch in 2011, WeChat has been known for its founder Zhang Xiaolong's "restraint," not changing WeChat itself for the sake of "trends." Even today, the experience of browsing Moments is fundamentally no different from when it was first introduced a decade ago.

"One of the things Zhang Xiaolong is most satisfied with about this feature is that once it was released, there was almost no room for improvement, and it has been running stably for ten years." This is how Zhang Peng, founder of Geek Park, summarized WeChat's product logic after a conversation with Zhang Xiaolong. This has not changed even after WeChat became a truly "national social app."

After the release of DeepSeek R1, these labels were suddenly torn off like masks by the WeChat team: On the evening of February 15, users began to notice that an "AI Search" feature integrated with DeepSeek R1 appeared in their WeChat search—this was less than a month after DeepSeek officially released and open-sourced R1 on January 20.

In the WeChat search interface, you can see the AI Search button | Image source: Geek Park

For users who received this update, by clicking into the search bar at the top of the WeChat homepage, they can see an "AI Search" button below their search history. By clicking in, they can see two response model options provided by this feature. The first is called "Quick Answer," and the second is the "Deep Thinking" mode powered by the open-source DeepSeek R1.

There are two model options, with the second being the version powered by DeepSeek R1 | Image source: Geek Park

It is worth mentioning that although this is a gray update received by only a portion of users, the speed of this update's rollout is much faster compared to recent gray updates of WeChat's features like Callkit. From the first leak to the large-scale appearance of the first batch of users receiving the push, it took less than six hours. This completely overturned the impression of WeChat's feature updates being "conservative" over the past few years This update is not tied to the version number of the software itself, so there is no need for you to update the WeChat version from the App Store or other application stores. The author found this entry after manually clearing the WeChat background on the phone and restarting it.

In other words, if your WeChat has not yet received this update, there is no need to worry too much— the DeepSeek R1 entry in your WeChat is actually hidden in the code, and the testing qualification may be coming soon.

From the open-source and acknowledgment statement accompanying the feature, it can be seen that the DeepSeek R1 built into WeChat is based on the open-source version, but it does not explicitly mention the model size used, whether it is the 671B "full-blooded" R1 version.

Open-source statement page for WeChat's "AI Search" feature | Image source: Geek Park

As we all know, the WeChat content ecosystem, including official accounts and video accounts, has long been an "island" independent of major search engines. This situation has not improved even after the emergence of connected language models like ChatGPT4o. Therefore, when testing the built-in R1 in WeChat, I was most excited about whether it could finally use the capabilities of large models to soar freely in the ocean of WeChat's content platform.

Unfortunately, in the current actual testing, the answer is negative.

For example, I tried feeding it a link to a WeChat official account article, but it could not retrieve relevant information from the WeChat platform. It could only search for related content on the internet outside the WeChat official platform based on relevant fields in the link, and sometimes it couldn't even recognize that this was a link from a WeChat official account article.

The current version of AI Search cannot recognize WeChat content links | Image source: Geek Park

In other questions related to official accounts, although AI Search can provide relatively accurate responses based on article searches and R1's reasoning capabilities, the sources of its content are closer to results derived from distributing official account content on other platforms, rather than citations from the WeChat content platform itself The responses regarding the WeChat platform content are more based on the integration of existing internet content | Image source: Geek Park

Currently, the built-in R1 in WeChat has not enhanced the retrieval generation (RAG) for content within the WeChat platform, nor optimized the output results.

This is both a significant shortcoming of the current beta version of WeChat AI search and one of the obvious important update directions for WeChat's integration of AI functions in the short term.

Overall, the current WeChat AI search experience is also very lightweight. It not only does not support continuous dialogue but also does not support uploading various file contents to assist in questioning and searching. Even after you exit the chat interface, the current dialogue memory content will be directly destroyed and not retained.

As a "national application," WeChat has integrated an entry within the app in less than a month since the official release and open-sourcing of DeepSeek R1, which is undoubtedly an exciting development.

The reason for the excitement is that what WeChat is doing is something only a few manufacturers can achieve.

After having a mobile app for dialogue generation models, seizing the entry on smartphone desktops has been something many AI applications, including ChatGPT and Perplexity, are doing by setting themselves as the default voice assistant, attempting to seize the "entry control" from smartphone brands as much as possible.

However, for a relatively closed platform like WeChat, it seems that currently, only the WeChat team can rely on WeChat's existing vast Chinese content ecosystem to do and do well in this regard.

From being "extremely conservative" to now becoming one of the first chat applications to join DeepSeek. This obvious change can only be reasonably explained by the fact that the WeChat team has recognized the enormous potential of the reasoning model represented by DeepSeek R1 in the application on the WeChat platform and has decided to act quickly to become the leader of this wave of change, ensuring that the WeChat user experience does not fall behind other competitors.

This is not the first time the WeChat team has taken action in the field of AI large models. The WeChat input method added the "one-click AI Q&A" feature last June to allow users to achieve content responses from language large models within the WeChat input method.

However, at that time, this feature was based on Tencent's own Hunyuan AI large model and could not serve as a text generation tool.

![](https://mmbiz-qpic.wscn.net/mmbiz_jpg/8cu01Kavc5ZEVtzI0g1e6yH8DzPoL0tFB2Pmg36fibDiajiae0sRcMqxJicJ6kMxor95VtmxQK5dp6n93yqAvn7dKA/640? The existing "Ask AI" feature in WeChat input method | Image source: Geek Park

In terms of experience, WeChat input method is suitable for "flash of inspiration" questions in various chats. The AI search within WeChat is likely to focus on the existing content ecosystem of WeChat, leveraging user chat content and platforms like WeChat Official Accounts to deeply explore application scenarios.

This kind of "distributed" AI capability experience is quite similar to Apple's product philosophy: in the Apple Intelligence capabilities released last year, Apple did not present a groundbreaking model that shocked the world, but instead chose to embed existing models like ChatGPT into various corners of the mobile ecosystem, including notes, photos, and input methods.

Apple's AI capabilities "scattered" across various applications | Image source: Geek Park

On the surface, Apple Intelligence may seem less intelligent compared to contemporaries like Google and OPPO, who are also pushing mobile AI applications. However, what Apple is actually doing is allowing AI to quietly integrate into users' lives, providing users with more usage scenarios through its capabilities.

From this perspective, although Tencent, which lacks a mobile operating system, missed one of the most important entry points in the mobile internet era, WeChat, as an important communication platform today, has become an indispensable part of the process of "truly popularizing AI."

In China, from centenarians to children who have just learned to use smartphones, everyone has a basic understanding of WeChat. These features are also the key to further "smoothing" AI capabilities and reducing the learning costs of AI.

For generative AI, which has already passed its explosive point, exploring the popularization of AI applications to enable AI capabilities to "undergo qualitative changes" through long-term use by more users is what we should truly look forward to in the future of WeChat AI.

We can even conclude that: The AI capabilities in WeChat may not be the most exciting, but they have the greatest opportunity to truly "change the world."

Author: Zhang Yongyi, Source: Geek Park, Original title: "After testing the DeepSeek R1 officially supported by WeChat, I understood how Tencent is betting on AI applications." Risk Warning and Disclaimer

The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investment based on this is at one's own risk