Flashspoter - Various generative AI services are again experiencing usage limit adjustments, especially for free users. In recent days, Google and OpenAI updated their daily usage policies for their video, text, and image models. This change had to be made after demand was significantly increased during the long holiday period, which led to a huge spike of computing loads and very few GPUs available for users. Their announcement is about how they are going to decrease the daily request number of non-subscribed users, whereas the limit that is set for paid subscribers will still be the same.
Google began the customization with the Nano Banana Pro, the latest image model released in conjunction with the launch of the Gemini 3 Pro. The system initially provided users with three image builds per day for FREE, similar to the previous scheme on the Gemini 2.5 Pro model. But now the limit has been reduced to two pictures per day. This information is contained in an internal Google Support document that later became publicly known. The company also confirms that usage limits are subject to change and will be updated daily. The statement reflects Google's stance that it is starting to be more cautious about regulating server load, especially after increased interest in image editing and Generation features.
In addition to the image model, free access to the Gemini 3 Pro is also undergoing revision. When it was introduced in mid-November, free users could get up to five daily requests. But now those guidelines are being replaced with the description “Basic access, daily limits may change frequently,” which marks the loss of a precise number for daily limits. Although Google did not elaborate on the details of its new numbers, this pattern of restrictions suggests a tightening consistent with the Nano Banana Pro Policy. Meanwhile, Google AI Pro customers keep getting up to 100 requests per day, and AI Ultra users get up to 500 requests per day with no limit changes.
On another front, OpenAI is tweaking Sora, its video generation model, which is currently one of the most demanded features in the AI ecosystem. The firm has imposed a limit of six video generations per day for free users. In contrast to the previous limits that were temporary, this update is a new setting that applies until further notice. OpenAI explains that video creation requests consume more computational power than text or image-based requests. Reducing the limit for free users is considered to be a way of helping to equalize access, maintain service stability, and lessen the pressure on GPUs that are currently operating at close to maximum capacity. The restrictions for paid users - such as ChatGPT Plus and Pro - remain unchanged, and if necessary, additional purchasing options are available.
At the same time, these two big companies tightening their measures simultaneously signals that the AI industry is moving towards a phase of infrastructural stabilization. After a full year of rolling out large models and increasingly complex multimedia features, the demand for computing is growing faster than the ability to expand the servers. Although the daily limit may be a source of disappointment for some free users, it also mirrors a business strategy adjustment where premium Access is increasingly becoming the main avenue for intensive use. But both companies still guarantee that free users will be able to test their services, only that the capacity will be limited.
As companies look for a balance between wide access and sustainable operations, this trend of limiting is going to continue to change. Due to the increasing demand and the relatively high cost of the computing infrastructure, the company will probably keep some sort of tiered plan which will gradually lead current users to have to choose paid plans. However, breakthroughs in AI model efficiency and improvements in server architecture may make it possible to have more stable free limits at some point.
Post a Comment