In February 2026, the Chinese technology review channel “Geek Bay” purchased 44 mobile phones out of its own pocket and released a video with a title that was so straightforward that it was almost provocative: “Mobile Game Performance Review: Manufacturers are so crazy about cheating!” 》. The video revealed that Chinese mobile phone manufacturers collectively tampered with “media test machines” and no one except the iPhone was spared. The film was blocked by the entire Internet less than a week after it was released, but it gave birth to China’s first official “Online Evaluation Activities Standards” just four months later.
China’s official “Geek Bay Terms” require that test prototypes must not be “special supply machines”
Geek Bay flipped the table: 44 mobile phones tested “all cheating”
Geek Bay has more than 5 million fans on Bilibili and nearly 600,000 subscribers on YouTube. It has become a practice to publish high-level reviews of mobile game performance during the Chinese New Year every year. The biggest difference between this issue, which was released on February 16, 2026, and previous years is that it no longer uses A, B, and C to anonymously refer to manufacturers, but directly names Xiaomi, Honor, vivo, OPPO and other brands.
The video points out that Chinese mobile phone manufacturers have a two-tier cheating mechanism. The first level is “special selection”: manufacturers will send test machines to the media to select chips with better performance. These chips have lower power consumption and better performance. Consumers cannot get the same level of products in the retail machines they buy on the market. The second level is “special adjustment”: the manufacturer has pre-embedded a whitelist in the system. When the phone detects that testing software (such as a benchmark tool) or a specific game (such as “Genshin Impact” and “Singing Tide”) is being executed, the power consumption limit will be automatically lifted and the performance will be fully turned on. The measured data will naturally be far better than the daily use experience.
The Geek Bay team spent money to purchase 44 commercially available mobile phones and conducted actual measurements on frame rate, power consumption, touch delay, heat and other items. The results showed that the iPhone was the only “honest” participant, and the performance of the newly purchased retail model was even better than that of the media machine previously sent for testing. On the other hand, the performance of all retail models of Chinese brand mobile phones was significantly behind the media machine test data, and “everyone was overturned.”
The video was blocked by the entire network: the road to silence from Bilibili to YouTube
The video was removed from Bilibili without warning on February 22. Geek Bay then uploaded the original video files to Baidu Cloud and authorized netizens to freely reprint them. However, the Baidu Cloud link was quickly blocked, and the backup videos that appeared on the site were also removed one after another.
Netizens turned to YouTube to watch, and a large amount of support poured into the comment area: “Like it, this is the evaluation team that consumers want.” “I like experimental tests that tell the truth.” However, on February 25, the video on YouTube was also set to private and has now been made public again, but comments are closed.
According to a discussion on the PTT MobileComm board, people on the other side revealed that the trigger for the video being removed was that after an executive from a mobile phone manufacturer saw the video, he privately contacted Chinese regulatory authorities to request a complete removal, and the executive did not discuss it with the company’s public relations department. After the situation escalated, various manufacturers were asked to re-examine the process of sending products to government units for testing.
极客湾03月28日直播片段揭露大横评视频下架原因
观众：方便聊一聊为什么横屏会被下吗？
云飞：其实原因没有你们想的那么复杂省流：某个厂的高管私自联系监管层，各厂商公关部门对大横评下架也很惊讶，因为谁把它搞下去都是双输局面，公关集体意志上不想动。 https://t.co/704n2TdwcP pic.twitter.com/ihApFb7maa
— 奶昔🥤 (@realNyarime) March 28, 2026
Four months later, the official took action: the “Online Assessment Activities Standards” were officially issued
On June 8, 2026, the Cyberspace Administration of China and the State Administration for Market Regulation jointly issued the “Specifications for online assessment activities》. This document is regarded by the outside world as a direct policy response to the Geek Bay incident.
The regulations clearly point out that in recent years, online evaluations “have had problems such as exaggerated publicity, only evaluation but not testing, and integration of business and testing”, which “not only affects consumer trust and shopping experience, but also disrupts the market environment.” In response to these chaos, the specification puts forward three core requirements:
- Samples must be from commercial sources: The samples selected for the evaluation should be “common commodities that consumers can purchase from the market and whose sources can be traced” and “should not be special items prepared for evaluation activities.”
- Interested relationships must be disclosed: If you accept entrustment or sponsorship from a third party or have an interest relationship with the relevant parties of the evaluation sample, you should make a “conspicuous reminder”
- Performance testing must be entrusted to a professional organization: When it comes to testing of product functions, performance and other items, an institution with statutory inspection and testing qualifications should be entrusted to perform it in accordance with relevant standards.
In addition, the specification also requires that “the product has not been tested and only subjective feelings are used to evaluate the product, and an explanation should be provided.” Website platforms need to strengthen the management of evaluation content and promptly accept and handle relevant complaints and reports. The policy issued by China’s official government this time seems to seek justice for Geek Bay, but it also contains a provision: “When engaging in online evaluation activities, it is not allowed to use different standards and different methods to compare similar products horizontally and vertically, and the content of the test report issued by the inspection and testing agency is not allowed to be taken out of context. “In order to enhance the reputation of the review sample by disparaging other products, you must not use insulting terms, falsely describe the product, or maliciously slander the product.” This is equivalent to the fact that this type of horizontal comparison review video of different brands of mobile phones is also explicitly prohibited, which means that there will no longer be such Dahengping videos.
From cases to systems: How a review channel leverages industry norms
Looking back at the timeline of the entire incident: the video was put on the shelves on February 16, removed from Station B on February 22, YouTube was closed on February 25, and the official regulations were issued on June 8. From a blocked review film to a national administrative standard, less than four months passed.
The clause “not to be special items prepared for evaluation activities” in the specification accurately corresponds to the “special selection” chaos exposed by Geek Bay; “acceptance of third-party commissions should provide a prominent reminder” refers to the gray area of ”integration of business and testing” that has existed for a long time but has never been expressly restricted.
For Taiwanese consumers, this incident also has reference value. Chinese mobile phone brands occupy a certain share of the Taiwan market. When there is a performance gap between the “media machines” and “retail machines” of these brands, overseas consumers are also the victims of information asymmetry. It remains to be seen whether the standards can be truly implemented and whether the evaluation data will return to reality.
Source: KOCPC Chinese