When large AI models are repriced: CloudWalk releases U2, ushering in the 'DeepSeek Moment'

The large model industry has long been immersed in a consensus that is almost taken for granted as correct.

Only with large parameters can the model be powerful; only with a sufficiently long context can the capabilities be comprehensive; only with a sufficiently complex reasoning chain can its intelligence level be demonstrated.

So in the past few years, from hundreds of billions of parameters to trillions of parameters, from hundreds of thousands of token contexts to millions of tokens, and from single responses to increasingly long reasoning, large model companies have continuously pushed the technological boundaries, and the capital market is also willing to pay for this stronger imagination. The model rankings change frequently, and the training costs are constantly being pushed up. GPUs have become one of the most expensive means of production.

However, when the passion for blindly piling up parameters fades, the industry begins to face an unavoidable and embarrassing reality: Dense models with thousands of billions or even trillions of parameters have simultaneously pushed up astronomical training and reasoning costs, erecting extremely high deployment thresholds.

Whether for high - flying enterprises or individual geeks, there is a huge gap between the ideal "emergence of intelligence" and the real - world situation of "unaffordable and difficult to use".

The more crazy the first half was, the more harsh the second half will be.

At this moment, some far - sighted players have realized that a paradigm shift is taking place, and generative AI is comprehensively evolving into productive AI.

It is necessary to integrate sufficiently hardcore intelligent capabilities into the real industrial capillaries smoothly with lower comprehensive costs and more stable delivery methods.

At this crucial juncture that determines the trend of the industry, an AI veteran has provided a solution for reconstructing efficiency through a hardcore upgrade of the base model.

Today, iFlytek's subsidiary, Unisound, officially released its new - generation general large - language model base, U2.

This is not only the most important technological iteration of the base model since Unisound's listing, but also a key milestone in its comprehensive transformation into an "Agentic large - model company".

While the industry was still chasing the chat - showy skills of generative AI, Unisound prospectively proposed the brand - new concept of "productive AI". The meaning behind it is very clear: The ultimate value of AI is not to generate content, but to solve complex tasks in the real world.

The proposal of this concept has made Unisound a well - deserved pioneer: While peers were still talking about "emergence", Unisound had already started thinking about "getting things done"; while others were competing in parameter scale, it had already seen through the commercial essence of intelligent density and token value. It is this advanced cognitive position that has allowed Unisound to gain the right to define the rules before the main competition of large models begins.

This enterprise, which has been struggling in the AI field for more than a decade, did not choose to join the blind parameter - consuming war. Instead, at this juncture, it used the new underlying logic represented by U2 to announce to the industry the recalculation of the commercial value of large models. Unisound has built an almost irreplicable moat with more than a decade of know - how in vertical scenarios, which is also the fundamental confidence for it to stay in the core position of the first echelon of domestic large models.

It's about sound, but not just sound

To understand the second half of the large - model era, one must first see clearly the players at the table.

In the domestic large - model camp where numerous players are vying for dominance, Unisound, founded in 2012, is a rather unique existence. It initially entered the market through voice recognition and was once active in scenarios such as smart healthcare, smart home, and in - vehicle cockpits. Over the past decade or more, it has experienced the complete technological cycle from statistical learning, deep learning to the large - model era. Therefore, it is often regarded as a somewhat "old - school" AI player.

For a long time in the past, because of the word "sound" in its name, the outside world habitually labeled it as a voice - recognition company. During the peak of the large - model craze, when the outside world's attention was attracted by the new Internet upstarts and the "Six Little Tigers" that had raised billions of dollars in financing and were in the spotlight, Unisound, which was in the silent period of its Hong Kong stock listing, seemed rather low - key.

Fortunately, the era of chatty generative AI ended in 2025. Everyone realized that productive AI that can get things done is what matters. At this moment, the industry suddenly found that Unisound's previously underestimated core business has instead become its broadest moat in the era of intelligent agents.

"Behind the sound is language, and behind the language is intention. What we listen to is not the sound, but the consciousness behind the sound." Huang Wei, the founder of Unisound, explained the "sound" in Unisound's name in this way.

In his understanding, there are always three levels of human - machine interaction: The first level is "understanding the speech", that is, voice recognition, which converts sound into text; the second level is "understanding the intention". When a user says "I'm cold", they don't just want a response, but hope that the air - conditioner will automatically adjust the temperature and the curtains will automatically close; the third level is to understand the deeper consciousness and scenarios - when an elderly person living alone casually says "I have nothing to do today", can the AI recognize the loneliness from the tone and pauses and actively trigger companionship or reminders?

From voice recognition, natural language understanding to today's large models and agents, Unisound has always been doing one thing: making machines truly understand humans and helping humans complete tasks.

In the real physical world of human - machine interaction, to make machines truly serve humans, a series of engineering problems such as multi - round interaction, long - chain tasks, complex environmental noise, and human - machine collaboration must be solved.

This scenario know - how and multi - round interaction engineering experience gained from struggling in serious and complex vertical scenarios are the natural soil for the survival of native agents.

Based on this profound insight, the newly released general large base U2 of Unisound is internally positioned as a native intelligent agent large model (Agent - Native Model). Its size, training goals, and optimization directions are all designed around "executing tasks".

In terms of the technical path, Unisound did not follow the industry - wide route of "completing model training and then attaching an agent framework". Instead, it proposed a more radical idea:

First, it is the native agent model + Harness collaborative evolution mechanism. In the past, most intelligent agent systems were more like putting a shell outside the general chat model - the model only focuses on speaking, and tasks such as planning, tool invocation, and task execution are all handed over to the external framework. In fact, the model itself does not really "understand" these things. However, U2 directly internalizes the complete capabilities of how to plan, execute, and accept the results into the model layer during the training stage. During the training process, the model and the Harness (task execution scaffolding) continuously evolve collaboratively: as the main structure of the model becomes more and more complex, the support nodes and verification accuracy of the scaffolding also extend and become more refined; and the more precise and strict scaffolding, in turn, ensures the solidity of each layer of the model's logic, forming a self - strengthening cycle.

Second, it is the systematic application of process supervision and curriculum learning. In order to make the agent as efficient as a person who can get things done cleanly, U2 introduced the "curriculum learning" method in the training process, allowing the model to progress step by step from easy to difficult, from short to long contexts, and from simple to complex tool invocations. In the trajectory of long - range tasks, U2 introduced an advanced process supervision method, using a better model to disassemble, evaluate, and correct each key node of task execution. U2 can not only see the final result but also optimize each intermediate execution path, achieving rapid convergence of learning.

Third, it is a more industry - level data ratio that is more inclined to serve the real economy and hardcore industries. While many large models still highly rely on general Internet corpora for generalization training, Unisound actively reduced the proportion of corpora from low - value scenarios such as entertainment and tilted more data resources towards high - value industry scenarios such as healthcare, medical insurance, insurance, government affairs, and industry. It also trained with real - scenario desensitized data accumulated from years of business implementation. It is worth mentioning that Unisound uses real - scenario desensitized data that has been long - term accumulated and is difficult to replicate in its business for synthesis and training, directly serving the real economy and hardcore industries.

After the reconstruction of the underlying capabilities, Unisound's U2 shows strong performance competitiveness without blindly piling up parameters. In evaluations such as IFBench for instruction following, U2's performance ranks among the top in the industry; in Claw - related evaluations, its agent and tool - invocation capabilities show strong advantages; in hardcore knowledge reasoning and long - context tasks such as GPQA, U2 also shows the ability to challenge the world's top large models; in GDPval, which measures the delivery ability for real - world office and knowledge work, U2 scored 72.5 points, demonstrating solid professional office capabilities.

Most importantly, U2 completely breaks the spell that "first - class performance must be tied to ultra - large parameters". It rejects parameter bloat and, through the ultimate MoE (Mixture of Experts) architecture and algorithm optimization, is committed to compressing the capabilities comparable to the world's first - class into a smaller parameter scale, pursuing strength in a small and cost - effective package.

This low - key and restrained AI veteran has entered the first echelon of domestic large models in a leading position.

How to close the business logic loop?

As a technology company with more than a decade of industry experience, Unisound knows better than the "newcomers" who have only entered the industry in recent years that while it is necessary to strive for technological generational leaps, the business logic cannot be ignored.

In the past, the large - model industry was used to discussing tokens from the single perspective of hardware and computing power. When the entire industry was competing to see who could generate more tokens and who had higher computing efficiency, Huang Wei, the founder of Unisound, calculated a more penetrating business account: "If the same one million tokens are produced, but they are all filled with idle chatter and nonsense, then no matter how high the computing efficiency is, there is no commercial value."

Based on this understanding, Unisound proposed a highly subversive business formula for the first time in the industry:

AI commercial value = intelligent density × token value.

Breaking it down, intelligent density means achieving a sufficiently high level of intelligence with smaller parameters and lower comprehensive resource input. Token value emphasizes that each invocation of the model must be directly convertible into measurable business results - either reducing risks or improving productivity.

The U2 model released today is the ultimate implementation carrier of this understanding and thinking. In order to make every penny of the customer's money count, U2 has achieved almost strict optimization in its underlying technology.

The previously mentioned Agent + Harness collaborative evolution mechanism is exactly to solve this problem. Through the co - evolution of the model and the tool - chain, U2 can complete task planning, tool invocation, execution, and acceptance with fewer interaction rounds, reducing the waste of tokens caused by a large number of trial - and - error processes and further improving the task completion rate.

At the same time, U2 uses a sparse Mixture of Experts (MoE) architecture at the bottom. Compared with traditional dense models that need to activate all parameters, MoE only activates a relevant part of the expert models for different tasks. According to the information disclosed by Unisound, U2 only activates about one - tenth of the parameters for calculation each time it processes a task, and the remaining parameters "sleep on demand". This means that the actual computational volume of the model during operation is much smaller than its full scale, significantly reducing the computing power cost required for reasoning while maintaining high performance.

What's more special is U2's redesigned thinking process. Some large models often expand a long reasoning process during complex reasoning - writing out the intermediate processes step by step. Although this method improves interpretability, it also brings another problem: users are paying for a large number of tokens that do not produce final value. U2 first conducts efficient exploration in the latent space, avoiding decoding each intermediate thought into visible tokens; when the task enters a critical stage, the model switches to explicit reasoning, completing logical calibration, process verification, and final decision - making through a readable and verifiable reasoning process. Unisound calls this "implicit thinking reasoning + explicit thinking verification".

"If these one million tokens are all idle chatter, no matter how high the efficiency is, there is little commercial value." Huang Wei once said.

This strategy of "pursuing high - value tokens with high intelligent density" quickly received real - money feedback in the commercialization battlefield, and the results demonstrated the infinite potential brought by the new formula.

The latest data shows that benefiting from the explosive growth in the demand for high - quality scenario tokens, Unisound's ARR of token invocation revenue in May increased by 600% month - on - month. And according to the current order momentum, it will continue to maintain a strong high - growth trend in June, with the expected ARR reaching 15 million US dollars.

In the second half of the large - model era, the ceiling of Unisound's business scale has been fully and completely lifted.

Behind this is an essential leap in Unisound's business model.

For a long time, traditional ToB companies have been deeply mired in the project - based quagmire - long delivery cycles, high customization levels, and hard - earned one - time money. However, with the release of the large model U2, through the continuous output of high - value tokens, Unisound's revenue model has been successfully and deeply bound to the intensity of customers' AI usage. As long as customers continuously invoke AI in their real business processes, the revenue will be like an open faucet, generating high - frequency, high - gross - margin repeat purchases.

Now, this efficient business closed - loop is being accelerated through Unisound's unique dual - wheel - drive layout:

In the To B segment (Beast Tooth Intelligent Agent Platform), Unisound uses U2 as the core base to make inroads in vertical industries. With an extremely high task completion rate, the company has recently won a series of bids in industry scenarios with strict accuracy requirements, such as healthcare, medical insurance, transportation, customer service, and employee badges. These high - value industries not only continuously contribute high customer unit prices but also use the real and high - quality business data accumulated in the scenarios to continuously feed back to the model base, forming a positive cycle of becoming smarter and having higher intelligent density with more use.

In the To C and developer segment (Public Cloud MaaS), Unisound relies on the OPC ecosystem to fully expand. Through the lower - threshold and higher - cost - performance model API invocation capabilities, it continuously and stably captures high - frequency token traffic and revenue from a wide range of independent developers and the C - end application ecosystem.

Instead of blindly following the limits of the Scaling Law, Unisound has found an ecological niche that perfectly matches its own resources and endowments. Unisound's hardcore technological self - hematopoietic ability, with a six - fold increase in monthly ARR, proves that in the second half of the large - model era, only players who can calculate the efficiency account clearly and hold the business closed - loop firmly have a truly infinite future.

The main competition has just begun

Looking back at the past year since its listing on the Hong Kong stock market, Unisound has presented a hardcore answer sheet that does not blindly follow the trend, is not divorced from reality, reconstructs efficiency with technology, and proves itself with business results. When the entire industry was deeply mired in parameter inflation and paying for expensive computing power experiments, this AI veteran, with its clear - headed strategic positioning and more than a decade of industrial in - depth cultivation, achieved technological foresight and found commercial value.

The release of the large - model base U2 is even a declaration to break the pattern. With the dual - high solution of "high intelligent density" and "