Following ChatGPT, Grok4 announced that it will be free and open. Is the AI free war about to begin? On August 11, xAI announced on the In this regard, paying users who pay between US$30 and US$300 per month directly complained of "subscription backstabbing".
According to actual testing by a reporter from "Kechuangban Daily", free users can only use Grok 4 for 3-5 query opportunities every 12 hours. After the limit is triggered, only Grok 3 can continue to be used. This restriction has triggered widespread doubts among users, pointing out that there is a strong contrast between the "loose usage restrictions" claimed by xAI and the actual experience.
5 free queries in 12 hours, paying users call them “backstabbed”
According to xAI, when users use automatic mode, Grok will automatically transfer complex queries to Grok4 for processing; while selecting "expert mode" can keep using Grok4 throughout the process.
A reporter from "Kechuangban Daily" confirmed the above functional settings through actual testing and operation of Grok software. In addition, the reporter found that when the "quick response" mode is enabled, it will switch to Grok3, while the "Heavy" mode is still only open to SuperGrok Heavy users.
It is reported that Grok has two subscription plans, namely SuperGrok at US$30/month and SuperGrok Heavy at US$300/month. Grok4, which previously required payment, is free, giving paying users a strong sense of being "stabbed in the back". Users believe that Grok4’s extensive freeing will weaken the value of its subscription service, and the reasonableness of monthly fees of US$30 and US$300 has also been questioned.
Some users directly questioned: "Why not upgrade the limited edition Grok4 Heavy to SuperGrok members? Then wouldn't I have wasted money on upgrading?"
In addition, regarding the daily limit for free use of Grok4, Grok's official account previously stated in a reply to netizens on the
A reporter from the "Kechuangban Daily" also found that when asking questions to Grok 4 for the fourth time, the system refused to answer the question, and a prompt popped up: "Your usage limit on Grok has been used up, please upgrade to SuperGrok." At this time, only the "quick response" mode equipped with Grok 3 can still work normally, while the "automatic mode" and "expert mode" are no longer available.
Many users have said that there is a big gap between xAI’s so-called “loose usage restrictions” and the actual number of uses provided.
A user confessed to a reporter from the Science and Technology Innovation Board Daily: "It is understandable that AI operating costs are high, and setting a limit is acceptable. However, the number of free times is really not enough. Sometimes a question is not fully understood, and the limit is used up after asking 5 times, and you have to wait another 12 hours before continuing. This mechanism is really unreasonable."
As of now, xAI and Musk have not yet given an official reply to users’ questions about “free and open”.
The free AI war is about to begin. Will you choose ChatGPT-5 or Grok4?
Grok 4 is fully free and open, just three days after the release of GPT-5.
Faced with OpenAI opening its flagship model to all free to Team, Plus, Pro, and Team users, this also breaks the previous practice of high-end AI models being only open to paying users. Therefore, some voices in the market believe that xAI’s move may be a strategy to follow up on the trends of industry competitors.
Although they both play the "free card", GPT-5 and Grok 4 have significant differences in their technical paths and functional characteristics.
Specifically, OpenAI’s upgrade to ChatGPT-5 focuses on breakthroughs in professional capabilities. For example, in the field of programming, it can efficiently generate visually coordinated website, application and game prototypes based on a single prompt; GPT-5 has also made great efforts in reliability, achieving a significant reduction in the hallucination rate, especially in the field of health and medical care. In-depth layout can generate personalized suggestions for users.
OpenAI CEO Sam Altman repeatedly emphasized at the press conference that GPT-5 can provide the experience of "dialogue with doctoral-level experts in various fields." Interestingly, Musk also made similar remarks when xAI released Grok 4 in July. Its promotion: "Grok 4 is better than the PhD level in all aspects."
Grok 4 exhibits significant technical features in terms of deep reasoning capabilities. According to the authoritative benchmark HLE (Human Ultimate Examination) data, its tool-free accuracy is 25.4%. If multi-agent collaboration is enabled, the score can be increased to 44.4%. This model has outstanding performance in the field of mathematics and has achieved perfect scores in the American Mathematics Invitational Competition AIME 25. In addition, its data processing architecture natively supports the integration of real-time information flows on the X platform, and has the ability to automatically capture social dynamics and event correlations.
The free strategies implemented by the two giants OpenAI and xAI have not only reshaped the value chain reconstruction and profit model innovation of AI services, but also pushed the battle for users in the AI model industry to a new height.
As the threshold for using AI continues to decrease, users’ expectations for functional depth and service experience will further increase. Under the wave of expanding demand for emotional value, AI companion services are moving from marginal needs to the core battlefield of the battle for users.
"Kechuangban Daily" reporter noticed that in the face of users' increasing emotional companionship and dependence on AI models, the two giants have also made corresponding strategic adjustments to attract more users and occupy the market high ground.
OpenAI has introduced four personality structures in ChatGPT-5, namely the picky (sharp tongue type), the mechanical (efficient and outspoken type), the listener (empathetic type), and the technical geek (passionate and inquiring type). Users can switch conversation agents according to their needs. In addition, in the face of strong market demand for the anthropomorphic characteristics of GPT-4o, OpenAI responded quickly and reintegrated the offline GPT-4o into the system.
On the other hand, xAI launches companion AI represented by the two-dimensional virtual image Ani, which transforms traditional tools into deep emotional companionship carriers through highly realistic personality modeling. Its core mechanism is to build an emotional attachment relationship, and ultimately achieve an upgrade in user stickiness through emotional value binding.
It should be noted that OpenAI CEO Sam Altman revealed some hidden concerns about users’ over-reliance on AI.
He said,"In the future, many people will truly trust the advice of ChatGPT when making their most important decisions. Although this prospect may be wonderful, it is also somewhat disturbing." Therefore, he believes that active thinking is needed to ensure that the overall benefits of AI far outweigh the disadvantages..