The paid version of Doubao is making substantial progress.

On June 11th, ByteDance’s large model application "Doubao" substantially accelerated its commercialization process.

Currently, Doubao is expanding the scope of gray testing for its professional paid subscription version and is expected to officially launch in late June. While maintaining basic free functions, the professional version precisely targets productivity scenarios, with the highest annual fee surpassing the 5,000 yuan mark.

As the leading AI application in China in terms of daily active users, Doubao’s commercialization pilot adopts a tiered pricing strategy.

According to information disclosed during the gray test phase, the pricing is consistent with what was first revealed in early May: the standard version monthly subscription is 68 yuan/month, annual subscription is 688 yuan/year; the enhanced version monthly subscription is 200 yuan/month, annual subscription is 2048 yuan/year; the professional version monthly subscription is 500 yuan/month, annual subscription is 5088 yuan/year.

The paid content mainly targets professionals with complex needs, covering high-value-added services such as software development, data analysis, professional design, process automation, and scientific research.

However, as the internal testing progresses, a mismatch in expectations between the high price and actual product experience is inevitably emerging.

Currently, a large number of users deeply involved in the gray test complain that "there’s little difference between the paid and the free version of Doubao."

According to observations by "China Entrepreneur" in user internal test groups, many testers found that in advanced productivity workflows, negative feedback mainly focuses on two points: first, the lack of personalized features; second, the overall experience is not optimal, and the paid version does not exhibit an overwhelming advantage over the free version.

Facing the gap in market experience, Doubao customer service responded to media questions saying that the difference between paid and free versions is indeed minimal, and the paid version only offers a bit more in terms of features.

The most essential stance is that the official team explicitly stated that generation quality depends on the underlying model itself, unrelated to whether the user has paid; payment does not equate to better output quality.

This response dispelled external expectations that Doubao’s professional version includes a hidden superior model.

From a business model perspective, Doubao’s current pricing strategy hasn’t adopted the internationally recognized approach of "the free version uses lower-tier models, the paid version exclusively accesses the latest high-tier models." Instead, it has shifted to a more pure "computing power channel fee" model.

After all, as the user base for large models expands, the consumption of GPU inference power for complex tasks increases exponentially. Charging high "quota fees" and "priority fees" to heavy, high-frequency users essentially commercializes the excess costs of computing power consumption.

However, in actual market dynamics, the pricing threshold above 5,000 yuan already far exceeds the price anchor of ChatGPT Plus (annual fee about 1,740 yuan RMB).

For domestic geeks and professionals, if paying a high premium only offers increased computing power usage and smoother experience during busy periods, but cannot provide decisive productivity gains in complex agent operations, advanced code generation, or deep logical inference, its actual conversion rate will face severe challenges.

It’s understood that when Doubao officially launches in late June, it may serve as a real test of domestic AI professional users’ willingness to pay.

However, the commercial closed loop for large models on both the C and Pro ends may ultimately still depend on generational leaps in model capabilities to support high premiums, not just tilting computing power quotas.

Risk Warning and DisclaimerThe market entails risks; investment needs caution. This article does not constitute individual investment advice and has not considered users’ specific investment goals, financial situation, or needs. Users should consider whether any opinions, views, or conclusions herein fit their particular circumstances. Investment based on this is at your own risk.