DeepSeek says that V4.1 Flash has gone through testing involving both internal teams and external evaluators. Based on those tests, the company claims the new Flash model has outperformed DeepSeek V4 Pro across multiple key metrics, making it a significant upgrade within the company’s model lineup.

V4 Pro Requests Will Move to V4.1 Flash

DeepSeek has also outlined how it plans to handle existing V4 Pro requests after the new model becomes available.

Once V4.1 Flash officially launches, but before the company releases V4.1 Pro, requests that would previously have been processed by V4 Pro will be automatically routed to V4.1 Flash. This pricing structure could make V4.1 Flash particularly attractive for developers and businesses running large volumes of AI workloads.

The move suggests that DeepSeek sees V4.1 Flash not simply as a faster alternative, but as a model capable of taking over a broader range of workloads previously handled by its Pro model.

DeepSeek Announces New Flash Pricing

Alongside the upcoming model launch, DeepSeek is also changing the pricing for its Flash series.

The new pricing is scheduled to take effect at 12:00 PM Beijing time on September 10, 2026. During off-peak periods, cached input will cost 0.02 yuan per million tokens, while uncached input will be priced at 1 yuan per million tokens. Output will cost 4 yuan per million tokens.

During peak hours, each of these rates will be twice the off-peak price.

This pricing structure could make V4.1 Flash particularly attractive for developers and businesses running large volumes of AI workloads, especially when response speed and API costs are both important considerations.

New Model Architecture and Native Multimodal Support

DeepSeek had already begun testing an intermediate version of V4.1 Flash before the official release.

According to a notice shared in one of DeepSeek’s official discussion groups, the test version uses a new model architecture and introduces native multimodal capabilities. The company also indicated that the model delivers stronger overall capabilities while responding faster and requiring lower operating costs.

Native multimodal support could be one of the more important changes in the V4.1 generation, potentially allowing the model to handle different types of information more naturally rather than focusing primarily on text-based tasks.

What to Expect From DeepSeek V4.1 Flash

The upcoming V4.1 Flash launch appears to be focused on improving the balance between AI performance, speed and cost.

Rather than relying solely on higher model capability, DeepSeek is positioning the new Flash model as a more efficient option for real-world workloads. Its planned replacement of V4 Pro requests also indicates that the company is confident in the model’s ability to handle demanding applications.

DeepSeek has not yet provided a complete breakdown of every benchmark or technical specification. More details should become available once V4.1 Flash officially goes live.

For now, the combination of a new architecture, native multimodal support, faster processing and lower pricing makes DeepSeek V4.1 Flash one of the company’s more notable model updates to watch in September 2026.

The post DeepSeek V4.1 Flash Set to Launch Around September 10, Reportedly Surpassing V4 Pro appeared first on IGeeKphone China Phone, Tablet PC, VR, RC Drone News, Reviews.

According to the latest AI model usage data from OpenRouter, global AI model usage reached approximately 113 trillion tokens from August 24 to August 30, representing a 21.11% increase from the previous week.

Chinese AI models accounted for 55.16 trillion tokens during the same period, up 36.26% week over week. By comparison, U.S. AI models recorded around 17.07 trillion tokens, despite posting a much faster weekly growth rate of 76.71%.

Even with that surge from U.S. models, China maintained a substantial lead in total weekly usage and remained ahead for the 18th consecutive week.

Chinese Models Sweep the Top Three

The latest ranking also showed the strength of China’s AI ecosystem at the individual-model level.

The top three models by weekly usage were all from China, with an anonymous model known as Ox Alpha, or the “Niu Lai” model, moving into first place.

Ox Alpha recorded approximately 15.7 trillion tokens of usage during the week, representing a 36% increase compared with the previous week.

The model originally attracted attention because of its unusually strong performance and the limited information available about its developer.

Ox Alpha launched on August 20 with a context window of approximately 1.048 million tokens and a maximum output length of around 131,000 tokens. It also supports text, image and video inputs, giving it native multimodal capabilities. During its preview period, users could access the model free of charge.

Those specifications, combined with its rapidly increasing usage, quickly made Ox Alpha one of the most discussed anonymous AI models in the industry.

Zhipu Confirms Ox Alpha’s Identity

The mystery surrounding Ox Alpha lasted only a few days.

On the evening of August 26, Chinese AI company Zhipu publicly confirmed that Ox Alpha was actually GLM-5.3-Flash, a new member of its GLM family.

The announcement effectively ended speculation over who was behind the rapidly rising model.

Zhipu also revealed that GLM-5.3-Flash is the first natively multimodal model in the GLM-5 family. Unlike models that rely on separate processing components for different media types, its architecture is designed to handle text, images and video as part of its core multimodal capabilities.

Another major selling point is its pricing.

Zhipu positioned GLM-5.3-Flash at approximately one-tenth the regular price of GLM-5.3, with a limited promotional discount bringing the effective price down to roughly one-twentieth of the original GLM-5.3 pricing.

The combination of lower inference costs, multimodal capabilities and high performance appears to have helped the model gain traction quickly.

GLM-5.3-Flash Rises Rapidly

The growth of GLM-5.3-Flash was particularly notable because it entered the ranking only after its recent launch.

During its first week on the platform, the model climbed to sixth place, recording approximately 6.16 trillion tokens in weekly usage.

That performance is significant considering how little time the model had been available compared with more established competitors.

Its rise also highlights a broader change in the AI model market.

Usage is increasingly influenced not only by benchmark performance, but also by inference cost, context length, multimodal support and developer accessibility. A capable model that is inexpensive to call can potentially generate enormous usage even without having the strongest consumer brand recognition.

China’s AI Usage Continues to Expand

The latest figures also reveal a broader trend beyond the success of a single model.

Chinese AI models generated 55.16 trillion tokens of weekly usage, more than three times the volume recorded by U.S. models at 17.07 trillion tokens.

Although U.S. models grew faster on a week-over-week basis, China’s overall lead remains substantial.

The 18-week streak suggests that Chinese models have established a strong position in global API and developer usage, particularly as more models become available through platforms such as OpenRouter.

At the same time, the rapid growth rate of U.S. models shows that the competitive landscape remains highly dynamic.

What Ox Alpha’s Rise Means

Ox Alpha’s journey from an anonymous model to a publicly identified GLM release is arguably the most interesting part of the latest ranking.

Its rapid rise demonstrates how quickly a new AI model can attract developers when it combines competitive capabilities with aggressive pricing.

The model’s support for long contexts and multiple input formats also reflects where the AI industry is heading. This reflects the broader shift toward multimodal AI models, which can work with text, images and video within a unified workflow.

For Zhipu, the success of GLM-5.3-Flash could provide another opportunity to expand the global reach of the GLM family.

For the wider AI market, however, the numbers point to an increasingly intense competition between Chinese and U.S. models.

China’s 18-week lead in total weekly token usage is notable, but the rapidly changing individual-model rankings show that the race for AI usage is far from settled.

The post China’s AI Models Lead Global Usage for 18 Straight Weeks as Ox Alpha Takes No. 1 Spot appeared first on IGeeKphone China Phone, Tablet PC, VR, RC Drone News, Reviews.

Unlike traditional AI models that rely on APIs or predefined commands, Qwen-UI-Agent can see and understand what’s displayed on a screen and perform actions accordingly. It can navigate apps, operate software, and execute tasks across different device types, effectively serving as a general-purpose digital device operator.

Benchmark Performance: Top-Tier Results Across the Board

According to Alibaba, Qwen-UI-Agent has achieved state-of-the-art or near-state-of-the-art results across multiple industry benchmarks:

Alibaba states that the model has reached or exceeded the performance of leading international models—including GPT-5.6 Sol, Claude Opus 4.8, and Gemini 3.1 Pro—while retaining the general-purpose and agent capabilities of the underlying Qwen foundation model.

Real-Device Training: Learning on Physical Hardware

A major differentiator for Qwen-UI-Agent is its emphasis on real-device training. Alibaba built a dedicated testing cluster containing:

• 100+ physical smartphones

• 150+ applications

This setup allows the model to move beyond simulated environments and learn how to perform tasks on actual hardware, making its behavior more reliable and practical for real-world deployment.

Built-In Security and User Control

Security is deeply integrated into the system. Qwen-UI-Agent can automatically reject high-risk requests, while sensitive actions—including payments, data deletion, and permission authorization—require explicit user confirmation before execution.

This dual-layer safety approach ensures that the agent remains helpful without becoming a liability.

CLI Integration and Batch Execution

The model supports a combination of GUI interactions and command-line interface (CLI) operations. It can also generate batches of actions simultaneously.

Alibaba notes that nearly half of all commands used during computer-based tasks can be handled through the command line, helping reduce unnecessary interaction steps and improve overall execution efficiency.

Continuous Learning and Scalable Training

Qwen-UI-Agent is trained through online reinforcement learning with trajectories exceeding 100 steps. More than 10,000 concurrent environments are used for training and iteration, while an AutoResearch-style AI-driven data pipeline continuously feeds new task information back into the development process.

This scalable training infrastructure allows the model to adapt and improve over time without manual intervention.

Cross-Device Coordination and DeepSearch Integration

Built on Alibaba’s Harness framework, Qwen-UI-Agent is designed to provide proactive assistance across devices. It can:

• Coordinate tasks between smartphones and computers

• Allow users to hand off an ongoing task between devices

• Work together with DeepSearch, combining web research with direct GUI interaction

This cross-device capability makes it a truly versatile assistant for complex, multi-step workflows.

Solving the API Gap: The Broader Vision

The broader goal of Qwen-UI-Agent is to tackle one of the biggest limitations facing today’s AI agents: software that does not provide an open API.

Instead of requiring developers to modify an application or build a dedicated integration, the AI can simply interpret the screen and interact with the software through its existing interface.

If successful, this approach could significantly expand the range of tasks AI agents can handle—bringing automated interaction to a much wider range of traditional applications and digital environments, without requiring any changes to the software itself.

Bottom Line

Qwen-UI-Agent marks a meaningful advance toward general-purpose AI capable of operating any digital interface. Its combination of strong benchmark results, real-device training, integrated security, and cross-platform coordination directly addresses the core shortcomings of today’s AI agents. By working with applications that lack open APIs, it opens the door to automation, accessibility, and digital assistance across a wide range of industries.

The post Alibaba’s Qwen-UI-Agent: The AI That Sees and Clicks on Every Screen appeared first on IGeeKphone China Phone, Tablet PC, VR, RC Drone News, Reviews.

The disclosure came via Anthropic’s “Risk Report for August 2026,” published on August 15. Buried within the document’s technical appendices was a notable benchmark comparison: Model 2 had edged out Claude Mythos 5 across multiple task categories, according to the company’s internal evaluation suite, CoBench v2.

But before the AI community declares a new king, the numbers tell a story of incremental progress rather than a paradigm shift.

What the CoBench v2 Benchmarks Actually Show

Anthropic’s proprietary CoBench v2 test suite measures AI performance across reasoning, coding, multilingual understanding, and safety compliance. While the full dataset wasn’t publicly released, the report’s summary indicated that Model 2 achieved consistently higher scores than Claude Mythos 5—but the improvements were described as “relatively small” and primarily contributed to overall comprehensive capabilities rather than any single breakout skill.

In plain terms: Model 2 is smarter, but not drastically so.

Source: Anthropic Risk Report – August 2026 (summary data)

The “Model 1” Mystery – What Does It Mean?

The report also made a passing reference to a “Model 1,” which has sparked immediate speculation across AI forums and social media.

The leading theory? Model 1 and Model 2 may represent successive iterations of the Claude Mythos Preview lineage—essentially, internal development checkpoints that Anthropic uses to track progress before rolling out public releases.

If that interpretation holds, Model 2 isn’t a competitor to Claude Mythos 5—it’s a preview of what’s coming next, possibly under a different product name when officially launched.

Why This Matters – Even If Gains Are Modest

At first glance, a 2–4% performance bump might not sound like headline news. But in the world of frontier AI, where models are already operating near the ceiling of available training data and compute, even single-digit gains represent meaningful engineering breakthroughs.

Here’s why:

• Efficiency over scale – Rather than simply throwing more parameters at the problem, Model 2 appears to achieve better results through architectural refinements.

• Safety improvements – The report emphasized that Model 2 maintained or improved alignment scores, a key priority for Anthropic’s “constitutional AI” approach.

• Competitive positioning – With OpenAI, Google DeepMind, and others racing toward GPT-5 and Gemini Ultra-class models, Anthropic’s ability to show consistent, measurable progress is a signal to investors and enterprise customers that it remains in the race.

What’s Next for Model 2?

Anthropic has not announced a public release date for Model 2, nor confirmed whether it will eventually be marketed as “Claude Mythos 6” or a completely new product line.

However, the August 2026 Risk Report serves as a de facto teaser—a controlled leak designed to manage expectations, gather feedback, and demonstrate transparency ahead of a broader rollout.

If past patterns hold, we could see an official announcement within 3–6 months, with enterprise API access arriving first, followed by consumer-facing features.

Final Takeaway

Anthropic’s Model 2 is real, it’s slightly better than Claude Mythos 5, and it’s coming—just not with fireworks.

For current Claude users, the news is reassuring: Anthropic is iterating, improving, and staying competitive. For the broader AI industry, it’s another data point in the slow, steady climb toward more capable, safer artificial intelligence.

And for the curious? The “Model 1” mention is a reminder that what we see today is only the latest checkpoint in a much longer development road.

The post Anthropic Model 2 Breakthrough? August 2026 Risk Report Shows Slim Lead Over Claude Mythos 5 appeared first on IGeeKphone China Phone, Tablet PC, VR, RC Drone News, Reviews.

Public suspicion is deeply rooted in Meta’s (and Zuckerberg’s) track record. He once promised that social media would “give everyone a voice” and connect the world. The reality, however, became a landscape of rage-baiting algorithms and ad-saturated feeds, not genuine connection. This persistent pattern of over-promising and under-delivering makes his latest “personal empowerment” pitch sound less like a sincere vision and more like another meticulously crafted corporate PR campaign.

In the AI race, Zuckerberg is attempting to play the “people’s champion” card, positioning himself as the optimistic counterpoint to rivals like Anthropic’s Dario Amodei, who emphasize risks. He argues that “we cannot afford to give an inch to China” and that accelerating development is the only way to truly empower individuals. Yet, the contradiction is glaring: his flagship on-device AI model, “Muse Glimmer,” can’t even run on a standard MacBook—it requires specialized hardware. This sharply undercuts his claim that the future is “for everyone.”

TechCrunch’s editors also highlight a deeper issue with such AI visions: they are often maddeningly abstract. Promises to “unleash creativity” or “cure cancer” sound inspiring, but when they come from a CEO whose business model is built on monetizing user data, they ring hollow. The public’s gut-level response is telling. When Sam Altman suggested AI could generate a podcast about your child’s interests for the drive to school, many retorted, “Why not just talk to your kid?” This instinctive recoil from technology replacing human interaction reveals a fundamental discomfort that no amount of polished prose can soothe.

Ultimately, Zuckerberg isn’t facing a simple PR misstep; he is confronting the culmination of years of eroded trust. When the account of credibility is overdrawn, even the most dazzling blueprint for the future struggles to convince a weary public to believe once more.

The post Mark Zuckerberg’s AI Utopia Vision: Why Skepticism Is Growing Over Meta’s Promises appeared first on IGeeKphone China Phone, Tablet PC, VR, RC Drone News, Reviews.

Amodei rejected the notion that his “doomsday predictions” were the primary driver of public panic. In his view, this is more like a long-overdue reckoning—a distrust of the tech industry that has been “decades in the making,” with AI serving as the final straw. “Ordinary people don’t trust companies, governments, or the tech industry and always suspect that we are cooking up some new way to screw them over.” This deep-seated skepticism, he argues, is the real backdrop against which the AI backlash is playing out.

Interestingly, Amodei handed his critics their strongest ammunition himself. He conceded that the most valid criticism of AI companies, including his own, isn’t about what they’ve said, but what they’ve failed to deliver. When promises like “AI will cure cancer” devolve into clichés, public patience inevitably wears thin. At its core, this is a credibility crisis born of overpromising and underdelivering.

On the regulatory front, Amodei struck a nuanced balance. He pushed back against the Silicon Valley orthodoxy that equates regulation with regulatory capture and concentrated power, while also acknowledging that AI is “structurally a technology that tends to concentrate power.” His solution? A carefully designed set of “rules of the road” that can simultaneously: (a) address AI’s cyber, bio, and alignment risks, (b) institutionally constrain the power of frontier AI companies, and (c) leave room for open-weights models while addressing their specific risks. This may be his attempt to chart a credible third path between the “doomer” and “booster” camps.

The post Anthropic CEO Explains Silicon Valley’s Trust Deficit: The Real Cause of AI Backlash appeared first on IGeeKphone China Phone, Tablet PC, VR, RC Drone News, Reviews.

According to Zhishu, GLM-5.3 delivers a major performance upgrade over the previous GLM-5.2, with improvements of around 50% across key evaluations. The company positions the new model as the strongest open-source model currently available, particularly highlighting its capabilities in software engineering, autonomous agents, terminal-based tasks, and other complex real-world workflows.

Zhishu also claims that GLM-5.3’s coding and agentic performance is approaching that of Claude Opus 5, although independent testing will be needed to verify the comparison.

GLM-5.3 Keeps the 740 Billion-Parameter Architecture

Despite the substantial performance improvement, GLM-5.3 does not reportedly increase the parameter count compared with GLM-5.2.

Both models are based on a 740-billion-parameter architecture, meaning Zhishu has not yet moved the GLM series into trillion-parameter territory.

Previous rumors had suggested that GLM-5.3 could potentially expand to more than one trillion parameters. That has not happened with the current release.

The larger parameter count could potentially be reserved for a future model such as GLM-5.5, although Zhishu has not officially confirmed such a roadmap.

Instead, GLM-5.3 appears to achieve its gains primarily through post-training and optimization, demonstrating that model performance can continue to improve substantially without simply increasing the underlying parameter count.

Zhishu Positions GLM-5.3 as a New Open-Source Leader

According to Zhishu, GLM-5.3 currently ranks at or near the top of several major open-source model evaluations.

The company says the model has made particularly strong progress in:

• Software engineering
• Complex coding tasks
• Terminal operations
• Autonomous agent workflows
• Cross-tool collaboration
• Long-running tasks
• Professional knowledge work

These capabilities are becoming increasingly important as AI models evolve from systems that simply generate text or code into agents capable of completing multi-step tasks with limited human intervention.

Rather than focusing exclusively on conversational quality, GLM-5.3 is being positioned as a model designed to actually perform work in real computing environments.

Major Gains on Terminal-Bench 3.0

One of the most dramatic improvements appears in Terminal-Bench 3.0, an evaluation designed to measure how effectively an AI model can complete complex tasks in real terminal environments.

According to Zhishu, GLM-5.3’s score increased from 4.6 with GLM-5.2 to 28.3.

That represents a substantial improvement and highlights the company’s emphasis on practical software-engineering capabilities.

Terminal-based tasks require models to go beyond generating a code snippet. They may need to inspect files, execute commands, troubleshoot errors, modify code, and iteratively work toward a desired result.

Strong performance in this category is therefore an important indicator of an AI system’s ability to operate as a practical software-development agent.

DeepSWE v1.1 Shows Stronger Software Engineering

GLM-5.3 also reportedly made significant progress on DeepSWE v1.1, an evaluation focused on longer-term software engineering and continuous code modification.

The model’s score reportedly increased from 46.2 to 66.9.

This improvement is particularly significant because real software development rarely consists of a single prompt followed by a finished answer.

Developers and AI coding agents often need to:

1. Understand an existing codebase.
2. Identify the source of a problem.
3. Modify multiple files.
4. Run tests.
5. Analyze failures.
6. Make additional changes.
7. Continue iterating until the task is completed.

The higher DeepSWE score suggests that GLM-5.3 has improved its ability to handle this type of extended engineering workflow.

Better Performance on Agents’ Last Exam

GLM-5.3 also reportedly improved its score on Agents’ Last Exam, an evaluation covering multiple professional scenarios and emphasizing cross-tool collaboration and longer-running tasks.

Its score increased from 23.8 to 28.5.

Agentic benchmarks are increasingly important because they measure capabilities that are difficult to capture through conventional question-and-answer evaluations.

A successful agent must be able to understand a goal, break it into steps, use external tools, respond to intermediate results, and continue working until the objective is achieved.

GLM-5.3’s improvement in this area suggests that Zhishu is attempting to build a model capable of functioning as a more autonomous digital worker rather than simply an advanced chatbot.

1,769 Score on GDPval-AA v2

Another notable result comes from GDPval-AA v2, an evaluation covering 44 occupations and designed to assess performance on high-value professional knowledge work.

Zhishu reports that GLM-5.3 achieved a score of 1,769 on this benchmark.

The result is notable because it suggests that the model’s improvements are not limited to programming.

While software engineering appears to be one of GLM-5.3’s strongest areas, the GDPval-AA v2 result indicates that the model is also being optimized for broader professional tasks.

This is an important direction for the AI industry as developers increasingly seek general-purpose agents capable of handling tasks across multiple knowledge-work domains.

GLM-5.3 vs GLM-5.2

The reported benchmark results show substantial improvements compared with the previous generation:

The figures above are based on Zhishu’s reported evaluation results. Independent replication will be important for determining how GLM-5.3 compares with other leading models under standardized testing conditions.

Coding and Agentic Performance Are the Main Focus

Zhishu’s messaging around GLM-5.3 makes it clear that coding and agentic capabilities are central to the new model.

The company claims that its programming performance is now approaching leading proprietary models such as Claude Opus 5 and that its coding ability surpasses other domestic Chinese models.

If independent evaluations confirm these claims, GLM-5.3 could become a significant competitor in the increasingly crowded AI coding-agent market.

The open-source nature of the model would make the development particularly noteworthy.

Developers, research organizations, and enterprises could potentially deploy, customize, and integrate the model into their own systems without depending entirely on a proprietary API.

Performance Gains Without a Larger Model

Perhaps the most interesting technical aspect of GLM-5.3 is that its improvements reportedly come without increasing the model’s parameter count.

Both GLM-5.2 and GLM-5.3 are said to contain 740 billion parameters.

This suggests that Zhishu has focused heavily on post-training techniques, data quality, reasoning optimization, tool-use training, and agent-specific capabilities.

The development reflects a broader trend in AI research: simply making models larger is no longer the only path toward improved performance.

Better training strategies can potentially extract considerably more capability from an existing architecture.

Could GLM-5.5 Move Into the Trillion-Parameter Range?

The absence of a trillion-parameter architecture in GLM-5.3 has also fueled speculation about future GLM models.

Some observers may interpret the current 740-billion-parameter architecture as a sign that Zhishu is deliberately reserving a major hardware and architecture upgrade for a future generation.

However, there is currently no official confirmation that GLM-5.5 will exceed one trillion parameters.

For now, GLM-5.3 demonstrates that Zhishu can achieve substantial performance gains without taking that step.

Why GLM-5.3 Matters for the Global AI Market

The release of GLM-5.3 comes as competition between open-source and proprietary AI models continues to intensify.

Companies around the world are developing increasingly capable models, while open-source projects are attempting to close the performance gap with commercial systems.

A model that combines strong coding performance, autonomous tool use, and professional knowledge-work capabilities could have applications across:

• Software development
• IT operations
• Automated testing
• Research
• Data analysis
• Business process automation
• Customer support
• Enterprise productivity
• AI agent platforms

The ability to execute tasks rather than simply generate responses is likely to become one of the most important competitive dimensions for the next generation of AI models.

Final Thoughts

The release of GLM-5.3 marks a significant step in Zhishu’s effort to compete at the highest level of the global AI industry.

Despite retaining the same 740-billion-parameter scale as GLM-5.2, the new model reportedly delivers major improvements in software engineering, terminal operations, agentic workflows, and professional knowledge work.

Its reported gains on Terminal-Bench 3.0, DeepSWE v1.1, and Agents’ Last Exam are particularly notable, while the 1,769 GDPval-AA v2 score points toward broader professional capabilities beyond programming.

Zhishu’s claim that GLM-5.3 is now the strongest open-source model and that its coding and agent performance is approaching leading proprietary systems is ambitious. Independent benchmarks and real-world testing will ultimately determine how it compares with the world’s top AI models.

Nevertheless, GLM-5.3 demonstrates an increasingly important trend: AI progress is no longer solely about making models larger. Through improved post-training and agent-oriented optimization, Zhishu is attempting to extract significantly greater practical capability from the same 740-billion-parameter foundation.

If these reported results hold up under independent evaluation, GLM-5.3 could become one of the most important open-source AI releases of 2026.

The post GLM-5.3 Released: Zhishu Claims New Open-Source AI Leader With Major Gains in Coding and Agentic Tasks appeared first on IGeeKphone China Phone, Tablet PC, VR, RC Drone News, Reviews.

It was an understandable description. OpenAI was entering one of the oldest and most entrenched layers of the internet with a browser built around AI agents rather than conventional tabs, bookmarks, and search. At the time, many observers believed the browser was finally ready for a major transformation.

The launch also triggered a rush among AI-browser startups. Several companies accelerated their own product releases, concerned that OpenAI’s enormous user base and brand recognition would otherwise overwhelm them.

But less than a year later, the story has taken a very different turn.

On August 9, 2026, just 292 days after Atlas launched, OpenAI discontinued the browser. Users were instructed to export their bookmarks before the shutdown, effectively marking the end of an experiment that was supposed to redefine how people interacted with the web.

The irony is difficult to miss: A company that set out to reinvent the browser ultimately had to teach its users how to migrate back to Chrome.

01. Chrome Never Really Had to Fight

The most striking part of the AI-browser story is that Chrome did not need a major counterattack.

Despite the arrival of numerous AI-powered browsers over the past year, Chrome’s global market share remained remarkably stable at roughly two-thirds. Depending on the measurement source, there were fluctuations, but the overall change remained relatively small.

Meanwhile, the combined market share of AI browsers remained below 1% in many browser statistics.

That gap is especially revealing when comparing the distribution of the underlying products.

ChatGPT had already reached a weekly active user base of hundreds of millions, giving OpenAI one of the world’s largest consumer software audiences. Yet Atlas struggled to convert that enormous audience into browser users at anything close to the same scale.

The problem was not necessarily that Atlas was technologically weak.

The problem was that the browser is one of the hardest consumer software categories to replace.

Switching browsers means moving bookmarks, saved passwords, extensions, payment information, browsing habits and countless small interactions that users have accumulated over years.

By comparison, adding an AI capability to an existing browser can be almost effortless.

The challenger asks users to move their digital lives.

The incumbent only needs to add a new feature.

That is an extremely uneven starting position.

The Cost of Breaking User Habits

Browser competition has never been purely about technical specifications.

For most people, a browser is infrastructure. Once a user becomes comfortable with a particular browser, there is little motivation to switch unless the alternative provides a dramatic improvement.

This is also why distribution has historically been more important than browser innovation itself.

Google’s long-running payments to secure default search placement on Apple devices illustrate how valuable the default entry point can be. The browser is not simply a piece of software; it is a gateway to the internet and a powerful distribution channel.

AI-browser startups therefore faced a fundamental problem.

They were attempting to replace a deeply embedded gateway rather than simply introduce a better AI assistant.

And the rise of mobile computing made that challenge even harder.

02. AI Browsers Arrived Without the Mobile Advantage

Another weakness of the AI-browser movement was its heavy focus on desktop computing.

Many of the products that attracted attention were initially designed around desktop workflows. Atlas itself remained a macOS-only product throughout its short life.

That created a significant strategic limitation.

The modern internet is no longer primarily a desktop experience. Smartphones account for a huge portion of global browsing activity, and mobile users have their own established browser ecosystems.

Trying to replace Chrome while failing to establish a meaningful presence across mobile platforms meant that AI browsers were effectively fighting over only part of the market.

In other words, they were trying to capture the browser category without controlling the places where much of the world’s browsing actually happens.

03. Big Tech Learned Faster Than Startups Could Innovate

Perhaps the most important lesson from the past year is that AI dramatically shortened the advantage normally enjoyed by innovative startups.

AI browsers introduced features such as:

• Sidebar AI assistants
• Web-page summaries
• Contextual question answering
• Automated browsing
• Agent-based task execution
• AI-powered research and navigation

These features initially looked like compelling reasons to switch browsers.

But the major browser companies quickly began incorporating similar capabilities into their existing products.

Google brought Gemini deeper into Chrome, while Microsoft expanded Copilot capabilities within Edge.

That fundamentally changed the competitive landscape.

A startup might spend months building an AI-powered browser experience. A platform company with an existing browser could potentially introduce a similar capability through a software update.

This undermined one of the traditional advantages of startups: the window of time between invention and imitation.

The classic “innovator’s dilemma” assumes that established companies may be reluctant to disrupt their existing businesses.

But AI browsers presented a different situation.

AI features did not necessarily threaten the core browser businesses of Google or Microsoft. As a result, incumbents had little reason to ignore the trend.

They could simply absorb the innovation.

And once the most attractive AI features became available inside existing browsers, the incentive to switch largely disappeared.

Chrome did not need to defeat the challengers.

It simply absorbed their best ideas.

04. The Browser Became a Feature, Not a Destination

That brings us to perhaps the most important conclusion from the collapse of the independent AI-browser boom.

The surviving products are generally not trying to make money from the browser alone.

Consider the different business models emerging around the category.

Perplexity’s Comet can function as an extension of its AI subscription ecosystem and a way to acquire users.

Dia, developed by The Browser Company and later acquired by Atlassian, can become part of a broader productivity and workplace software strategy.

Opera Neon has positioned itself as a premium AI browser with a subscription model, while smaller products can depend heavily on venture capital to finance development.

The common denominator is clear:

The browser is rarely the business itself.

Instead, it is becoming a distribution mechanism for another business.

It can acquire customers, provide access to AI services, extend a productivity platform, or serve as another interface for an existing ecosystem.

This represents a fundamental challenge to the idea of the independent AI-browser company.

The browser category will survive.

But the idea that dozens of standalone companies can build businesses simply by creating “a better AI browser” looks increasingly difficult to sustain.

05. The Chinese Market Offers a Similar Lesson

There is also a precedent for this development in China.

Rather than creating entirely new browsers specifically for the AI era, major technology companies have generally taken existing products and gradually integrated AI capabilities into them.

Alibaba has continued evolving its existing consumer software toward AI-powered work and productivity experiences. Tencent has integrated AI into established products such as QQ Browser, while 360 has incorporated AI capabilities into its security-focused browser ecosystem.

The pattern is consistent:

Existing products become AI products.

The browser remains the shell, while AI becomes the capability inside it.

That is almost the opposite of the strategy pursued by many independent AI-browser startups in 2025.

06. OpenAI Didn’t Abandon the Browser’s Most Important Idea

Atlas’s shutdown does not necessarily mean OpenAI has abandoned browser-based AI.

In fact, the company’s next strategy may be more important than Atlas itself.

Rather than maintaining a separate browser, OpenAI can take the most valuable capabilities developed for Atlas and distribute them through other products.

Some of Atlas’s functionality has reportedly moved into the desktop ChatGPT experience, including browser automation capabilities that allow AI agents to interact with websites on a user’s behalf.

Other capabilities can be delivered through a Chrome extension.

This represents a fundamental change in strategy.

OpenAI no longer needs to convince hundreds of millions of people to abandon their existing browser.

It can simply bring its agent into the browser they already use.

And that could be a much more powerful approach.

07. The Real Browser War Is Becoming an Agent War

The first generation of AI browsers asked a relatively simple question:

What if we built a browser around AI from the beginning?

The next generation is asking something much more consequential:

What if AI could operate the browser for you?

That distinction matters.

The future competition may not be about which browser has the smartest sidebar or the most attractive new-tab page.

Instead, it may be about which AI agent can successfully perform tasks across the web.

Imagine asking an agent to research several products, compare prices, complete a form, organize information, book an appointment, monitor a website, or perform a series of repetitive online tasks.

In that world, the browser itself becomes infrastructure.

The user does not necessarily care which browser is underneath the agent.

They care whether the agent can get the job done.

That puts OpenAI and Google back into direct competition—but on a very different battlefield.

The weapon is no longer the browser.

It is the AI agent.

08. What Happened to the Other AI Browsers?

Atlas was not the only casualty of the AI-browser boom.

Over roughly the same period, several other major developments reshaped the category.

Arc, developed by The Browser Company, stopped receiving active development as the company shifted its attention toward its AI-focused successor, Dia.

Sidekick was absorbed by Perplexity, with the product subsequently discontinued.

The Browser Company itself was acquired by Atlassian in a deal reportedly valued at around $610 million, turning what began as an independent browser company into part of a much larger enterprise software organization.

These outcomes are different on the surface—shutdown, acquisition, absorption and strategic repositioning—but they point toward the same underlying trend.

The independent AI browser is becoming increasingly difficult to justify as a standalone business.

The technology itself may survive.

The standalone category may not.

09. 292 Days That Changed the Strategy

Atlas existed for only 292 days.

That is an extraordinarily short lifespan for a product positioned as a fundamental reinvention of the web browser.

Yet those 292 days may ultimately prove useful for OpenAI.

The company learned that its enormous ChatGPT audience does not automatically translate into willingness to change a fundamental piece of personal computing infrastructure.

It also demonstrated that users are willing to adopt AI capabilities without necessarily adopting a new browser.

That distinction could shape the next phase of AI development.

The battle for the internet’s entry point is not over.

It has simply moved above the browser.

The Real Lesson: Don’t Make Users Move

The biggest lesson from the rise and fall of Atlas is surprisingly simple.

Users don’t necessarily want a new place to browse the web. They want their existing tools to become more capable.

For years, technology companies have competed by trying to persuade users to move from one platform to another.

AI changes that equation.

The most powerful AI products may not be the ones that force users to adopt a completely new environment. They may be the ones that quietly integrate themselves into environments users already understand.

That is why a Chrome extension can ultimately be more strategically valuable than a new browser.

And it explains the irony of Atlas’s short life.

OpenAI spent almost a year building a browser intended to replace the existing internet gateway.

In the end, its most valuable technology may live inside the gateway it was originally trying to replace.

The “Chrome killer” did not kill Chrome.

Instead, it helped demonstrate why the future of AI may not require killing the browser at all.

The next battle is not about replacing the browser. It is about making the browser intelligent enough that users no longer have to think about it.

The post OpenAI’s Atlas Shutdown Shows Why AI Doesn’t Need a New Browser appeared first on IGeeKphone China Phone, Tablet PC, VR, RC Drone News, Reviews.

The latest claims originate from information shared by X (formerly Twitter) user @synthwavedd, who stated that the project is internally known by the codename “mewfour.” While OpenAI has not officially confirmed the existence or launch schedule of Astra, the reports have fueled growing anticipation within the AI community.

A New Flagship Pre-Trained Model

According to the leaked information, Astra is an entirely new pre-trained foundation model rather than an incremental update to an existing system.

The source claims that Astra represents the largest model OpenAI has trained since GPT-4.5, suggesting significant increases in model scale and computational capability. Although no official technical specifications have been disclosed, the project is expected to build upon OpenAI’s recent advances in reasoning, multimodal understanding, and autonomous AI workflows.

Reportedly Solved Multiple Open Mathematical Problems

Interest in Astra intensified following earlier reports that an internal version of the model successfully solved 10 major open mathematical problems during testing.

According to previous reports, the computational resources required for those solutions were substantial. Based on estimated usage costs for OpenAI’s Sol API, generating the solutions reportedly consumed approximately US$2,000 worth of tokens.

Because these claims have not been independently verified or confirmed by OpenAI, they should be treated as preliminary until official technical details become available.

Designed for Complex Multi-Agent Workflows

Beyond raw reasoning performance, Astra is also rumored to introduce enhanced capabilities for coordinating multiple AI agents.

According to reports from BleepingComputer, the model is designed to:

• Coordinate multiple AI agents working together
• Handle long-running and complex tasks
• Support advanced reasoning workflows
• Improve collaborative AI operations
• Execute sophisticated multi-step problem-solving processes

If accurate, these capabilities would align with a broader industry trend toward agentic AI, where multiple specialized AI systems collaborate to complete tasks that extend beyond traditional chatbot interactions.

Focus on Long-Term Task Execution

Unlike conventional conversational AI models that primarily respond to individual prompts, Astra is reportedly optimized for sustained workflows involving multiple stages of planning, execution, and reasoning.

Potential applications could include:

• Scientific research
• Software development
• Enterprise automation
• Large-scale data analysis
• Engineering workflows
• Knowledge management
• Autonomous research assistants

Such capabilities would represent another step toward AI systems capable of acting as long-term digital collaborators rather than simple question-and-answer assistants.

Growing Competition in Frontier AI

The rumored launch comes amid intense competition among leading AI developers, including Anthropic, Google, Meta, xAI, and OpenAI, all of which are investing heavily in larger foundation models, advanced reasoning systems, and autonomous AI agents.

As organizations increasingly seek AI solutions capable of handling complex professional tasks, frontier models are evolving beyond conversational interfaces toward systems that can independently plan, coordinate, and execute sophisticated workflows.

Official Confirmation Still Pending

Despite growing speculation, OpenAI has not officially announced Astra, its release date, or any technical specifications. As with all pre-release information, the reported launch timeline and capabilities should be regarded as rumors until confirmed by the company.

If Astra does debut in the coming days, it could provide valuable insight into OpenAI’s next generation of foundation models and its broader vision for advanced AI systems capable of tackling increasingly complex reasoning and collaborative tasks.

For now, the AI community will be watching closely for any official announcements that clarify whether Astra will indeed become OpenAI’s next flagship model.

The post OpenAI Reportedly Preparing to Launch “Astra” Next Week, Rumored to Be Its Largest Model Since GPT-4.5 appeared first on IGeeKphone China Phone, Tablet PC, VR, RC Drone News, Reviews.

The post DeepGrove Unveils Maple-Preview-20B-A1B, a High-Speed On-Device AI Model for iPhone and Apple Silicon appeared first on IGeeKphone China Phone, Tablet PC, VR, RC Drone News, Reviews.