Many organizations would be hesitant to overhaul their tech stack and start from scratch.

Not Notion.

For the 3.0 version of its productivity software (released in September), the company didn’t hesitate to rebuild from the ground up; they recognized that it was necessary, in fact, to support agentic AI at enterprise scale.

Whereas traditional AI-powered workflows involve explicit, step-by-step instructions based on few-shot learning, AI agents powered by advanced reasoning models are thoughtful about tool definition, can identify and comprehend what tools they have at their disposal and plan next steps.

“Rather than trying to retrofit into what we were building, we wanted to play to the strengths of reasoning models,” Sarah Sachs, Notion’s head of AI modeling, told VentureBeat. “We've rebuilt a new architecture because workflows are different from agents.”

Re-orchestrating so models can work autonomously

Notion has been adopted by 94% of Forbes AI 50 companies, has 100 million total users and counts among its customers OpenAI, Cursor, Figma, Ramp and Vercel.

In a rapidly evolving AI landscape, the company identified the need to move beyond simpler, task-based workflows to goal-oriented reasoning systems that allow agents to autonomously select, orchestrate, and execute tools across connected environments.

Very quickly, reasoning models have become “far better” at learning to use tools and follow chain-of-thought (CoT) instructions, Sachs noted. This allows them to be “far more independent” and make multiple decisions within one agentic workflow. “We rebuilt our AI system to play to that," she said.

From an engineering perspective, this meant replacing rigid prompt-based flows with a unified orchestration model, Sachs explained. This core model is supported by modular sub-agents that search Notion and the web, query and add to databases and edit content.

Each agent uses tools contextually; for instance, they can decide whether to search Notion itself, or another platform like Slack. The model will perform successive searches until the relevant information is found. It can then, for instance, convert notes into proposals, create follow-up messages, track tasks, and spot and make updates in knowledge bases.

In Notion 2.0, the team focused on having AI perform specific tasks, which required them to “think exhaustively” about how to prompt the model, Sachs noted. However, with version 3.0, users can assign tasks to agents, and agents can actually take action and perform multiple tasks concurrently.

“We reorchestrated it to be self-selecting on the tools, rather than few-shotting, which is explicitly prompting how to go through all these different scenarios,” Sachs explained. The aim is to ensure everything interfaces with AI and that “anything you can do, your Notion agent can do.”

Bifurcating to isolate hallucinations

Notion’s philosophy of “better, faster, cheaper,” drives a continuous iteration cycle that balances latency and accuracy through fine-tuned vector embeddings and elastic search optimization. Sachs’ team employs a rigorous evaluation framework that combines deterministic tests, vernacular optimization, human-annotated data and LLMs-as-a-judge, with model-based scoring identifying discrepancies and inaccuracies.

“By bifurcating the evaluation, we're able to identify where the problems come from, and that helps us isolate unnecessary hallucinations,” Sachs explained. Further, making the architecture itself simpler means it’s easier to make changes as models and techniques evolve.

“We optimize latency and parallel thinking as much as possible,” which leads to “way better accuracy,” Sachs noted. Models are grounded in data from the web and the Notion connected workspace.

Ultimately, Sachs reported, the investment in rebuilding its architecture has already provided Notion returns in terms of capability and faster rate of change.

She added, “We are fully open to rebuilding it again, when the next breakthrough happens, if we have to.”

Understanding contextual latency

When building and fine-tuning models, it’s important to understand that latency is subjective: AI must provide the most relevant information, not necessarily the most, at the cost of speed.

“You'd be surprised at the different ways customers are willing to wait for things and not wait for things,” Sachs said. It makes for an interesting experiment: How slow can you go before people abandon the model?

With pure navigational search, for instance, users may not be as patient; they want answers near-immediately. “If you ask, ‘What's two plus two,’ you don't want to wait for your agent to be searching everywhere in Slack and JIRA,” Sachs pointed out.

But the longer the time it's given, the more exhaustive a reasoning agent can be. For instance, Notion can perform 20 minutes of autonomous work across hundreds of websites, files and other materials. In these instances, users are more willing to wait, Sachs explained; they allow the model to execute in the background while they attend to other tasks.

“It's a product question,” said Sachs. “How do we set user expectations from the UI? How do we ascertain user expectations on latency?”

Notion is its biggest user

Notion understands the importance of using its own product — in fact, its employees are among its biggest power users.

Sachs explained that teams have active sandboxes that generate training and evaluation data, as well as a “really active” thumbs-up-thumbs-down user feedback loop. Users aren’t shy about saying what they think should be improved or features they’d like to see.

Sachs emphasized that when a user thumbs down an interaction, they are explicitly giving permission to a human annotator to analyze that interaction in a way that de-anonymizes them as much as possible.

“We are using our own tool as a company all day, every day, and so we get really fast feedback loops,” said Sachs. “We’re really dogfooding our own product.”

That said, it’s their own product they’re building, Sachs noted, so they understand that they may have goggles on when it comes to quality and functionality. To balance this out, Notion has trusted "very AI-savvy" design partners who are granted early access to new capabilities and provide important feedback.

Sachs emphasized that this is just as important as internal prototyping.

“We're all about experimenting in the open, I think you get much richer feedback,” said Sachs. “Because at the end of the day, if we just look at how Notion uses Notion, we're not really giving the best experience to our customers.”

Just as importantly, continuous internal testing allows teams to evaluate progressions and make sure models aren't regressing (when accuracy and performance degrades over time). "Everything you're doing stays faithful," Sachs explained. "You know that your latency is within bounds."

Many companies make the mistake of focusing too intensely on retroactively-focused evans; this makes it difficult for them to understand how or where they're improving, Sachs pointed out. Notion considers evals as a "litmus test" of development and forward-looking progression and evals of observability and regression proofing.

“I think a big mistake a lot of companies make is conflating the two,” said Sachs. “We use them for both purposes; we think about them really differently.”

Takeaways from Notion's journey

For enterprises, Notion can serve as a blueprint for how to responsibly and dynamically operationalize agentic AI in a connected, permissioned enterprise workspace.

Sach’s takeaways for other tech leaders:

• Don’t be afraid to rebuild when foundational capabilities change; Notion fully re-engineered its architecture to align with reasoning-based models.

• Treat latency as contextual: Optimize per use case, rather than universally.

• Ground all outputs in trustworthy, curated enterprise data to ensure accuracy and trust.

  She advised: “Be willing to make the hard decisions. Be willing to sit at the top of the frontier, so to speak, on what you're developing to build the best product you can for your customers.”

In a packed theater at Fort Mason, after a whirlwind keynote of product announcements, OpenAI CEO Sam Altman sat down with Sir Jony Ive, the legendary designer behind Apple's most iconic products. The conversation, held exclusively for the 1,500 developers in attendance and not part of the public livestream, offered the clearest glimpse yet into the philosophy and ambition behind their secretive collaboration to build a new "family" of AI-powered devices.

The partnership, solidified by OpenAI's staggering $6.5 billion acquisition of Ive's hardware startup Io in May, has been the subject of intense speculation.While concrete product details remained under wraps, the discussion pivoted away from specifications and toward a profound, almost therapeutic mission: to fix our broken relationship with technology.

For nearly 45 minutes, Ive, in his signature thoughtful cadence, articulated a vision that feels like both a continuation of and a repentance for his life's work. The man who designed the iPhone, a device that arguably defined the modern era of personal computing, is now on a quest to cure the very anxieties it helped create.

Jony Ive's post-Apple mission, clarified by ChatGPT

The collaboration, Ive explained, was years in the making, but it was the launch of ChatGPT that provided a sudden, clarifying purpose for his post-Apple design collective, LoveFrom.

"With the launch of ChatGPT, it felt like our purpose for the last six years became clear," Ive said. "We were starting to develop some ideas for an interface based on the capability of the technology these guys were developing... I've never in my career come across anything vaguely like the affordance, like the capability that we're now starting to sense."

This capability, he argued, demands a fundamental rethinking of the devices we use, which he described as "legacy products" from a bygone era. The core motivation, he stressed, is not about corporate agendas but about a sense of duty to humanity.

"The reason we're doing this is we love our species and we want to be useful," Ive said. "We think that humanity deserves much better than humanity generally is given."

An 'obscene understatement': Jony Ive's quest to cure our tech anxiety

The most striking theme of the conversation was Ive's candid critique of the current state of technology — the very ecosystem he was instrumental in building. He described our current dynamic with our devices as deeply flawed, a problem he now sees AI as the solution to, not an extension of.

"I don't think we have an easy relationship with our technology at the moment," Ive began, before adding, "When I said we have an uncomfortable relationship with our technology, I mean, that's the most obscene understatement."

Instead of chasing productivity, the primary goal for this new family of devices is emotional well-being. It's a radical departure from the efficiency-obsessed ethos that dominates Silicon Valley.

When asked about his ambitions for the new devices, Ive prioritized emotional well-being over simple productivity. "I know I should care about productivity, and I do," he said, but his ultimate goal is that the tools "make us happy and fulfilled, and more peaceful and less anxious, and less disconnected."

He framed it as a chance to reject the current, fraught relationship people have with their technology. "We have a chance to... absolutely change the situation that we find ourselves in," he stated. "We don't accept this has to be the norm."

Buried in brilliance: why '15 to 20 compelling ideas' have become Ive's biggest challenge

While the vision is clear, the path is fraught with challenges. Reports have surfaced about technical hurdles and philosophical debates delaying the project. Ive himself gave voice to this struggle, admitting the sheer pace of AI's progress has been overwhelming. The rapid advancement has generated a torrent of possibilities, making the crucial act of focusing incredibly difficult.

"The momentum is so extraordinary... it has led us to generate 15 to 20 really compelling product ideas. And the challenge is trying to focus," Ive confessed."I used to be good at that, and I've lost some confidence, because the choices are, it'll be easy if you really knew there were three good ones... it's just not like that."

This admission provides context to reports that the team is grappling with unresolved issues around the device's "personality" and computing infrastructure. The goal, according to one source, is to create an AI companion that is "accessible but not intrusive," avoiding the pitfalls of a "weird AI girlfriend."

Beyond the screen: Ive's design philosophy for an 'inevitable' AI device

While no devices were shown, the conversation and prior reports offer clues. The project involves a "family of devices," not a single gadget.It will likely be a departure from the screen-centric world we inhabit. Reports suggest a "palm-sized device without a screen" that relies on cameras and microphones to perceive its environment.

Ive argued that it would be "absurd" to assume that today's breathtaking AI technology should be delivered through "products that are decades old." The goal is to create something that feels entirely new, yet completely natural.

"It should seem inevitable. It should seem obvious, as if there wasn't possibly another rational solution to the problem," Ive said, echoing a design philosophy often attributed to his time with Steve Jobs.

He also spoke of bringing a sense of joy and whimsy back to technology, pushing back against a culture he feels has become overly serious.

"In terms of the interfaces we design, if we can't smile honestly, if it's just another deeply serious sort of exclusive thing, I think that would do us all a huge disservice," he remarked.

The chat concluded without a product reveal, leaving the audience with a philosophical blueprint rather than a technical one. The central narrative is clear: Jony Ive, the designer who put a screen in every pocket, is now betting on a screenless future, powered by OpenAI's formidable intelligence, to make us all a little less anxious and a little more human.

Are data center architects ready to rethink their approach and embrace smarter, network-driven strategies to meet the demands of the future?

In a packed hall at Fort Mason Center in San Francisco, against a backdrop of the Golden Gate Bridge, OpenAI CEO Sam Altman laid out a bold vision to remake the digital world. The company that brought generative AI to the mainstream with a simple chatbot is now building the foundations for its next act: a comprehensive computing platform designed to move beyond the screen and browser, with legendary designer Jony Ive enlisted to help shape its physical form.

At its third annual DevDay, OpenAI unveiled a suite of tools that signals a strategic pivot from a model provider to a full-fledged ecosystem. The message was clear: the era of simply asking an AI questions is over. The future is about commanding AI to perform complex tasks, build software autonomously, and live inside every application, a transition Altman framed as moving from "systems that you can ask anything to, to systems that you can ask to do anything for you." 

The day’s announcements were a three-pronged assault on the status quo, targeting how users interact with software, how developers build it, and how businesses deploy intelligent agents. But it was the sessions held behind closed doors, away from the public livestream, that revealed the true scope of OpenAI’s ambition — a future that includes new hardware, a relentless pursuit of computational power, and a philosophical quest to redefine our relationship with technology.

From chatbot to operating system: The new 'App Store'

The centerpiece of the public-facing keynote was the transformation of ChatGPT itself. With the new Apps SDK, OpenAI is turning its wildly popular chatbot into a dynamic, interactive platform, effectively an operating system where developers can build and distribute their own applications.

“Today, we're going to open up ChatGPT for developers to build real apps inside of ChatGPT,” Altman announced during the keynote presentation to applause. “This will enable a new generation of apps that are interactive, adaptive and personalized, that you can chat with.”

Live demonstrations showcased apps from partners like Coursera, Canva, and Zillow running seamlessly within a chat conversation. A user could watch a machine learning lecture, ask ChatGPT to explain a concept in real-time, and then use Canva to generate a poster based on the conversation, all without leaving the chat interface. The apps can render rich, interactive UIs, even going full-screen to offer a complete experience, like exploring a Zillow map of homes.

For developers, this represents a powerful new distribution channel. “When you build with the Apps SDK, your apps can reach hundreds of millions of chat users,” Altman said, highlighting a direct path to a massive user base that has grown to over 800 million weekly active users. 

In a private press conference later, Nick Turley, head of ChatGPT, elaborated on the grander vision. "We never meant to build a chatbot," he stated. "When we set out to make ChatGPT, we meant to build a super assistant and we got a little sidetracked. And one of the tragedies of getting a little sidetracked is that we built a great chatbot, but we are the first ones to say that not all software needs to be a chatbot, not all interaction with the commercial world needs to be a chatbot."

Turley emphasized that while OpenAI is excited about natural language interfaces, "the interface really needs to evolve, which is why you see so much UI in the demos today. In fact, you can even go full screen and chat is in the background." He described a future where users might "start your day in ChatGPT, just because it kind of has become the de facto entry point into the commercial web and into a lot of software," but clarified that "our incentive is not to keep you in. Our product is to allow other people to build amazing businesses on top and to evolve the form factor of software."

The rise of the agents: Building the 'do anything' AI

If apps are about bringing the world into ChatGPT, the new "Agent Kit" is about sending AI out into the world to get things done. OpenAI is providing a complete "set of building blocks... to help you take agents from prototype to production," Altman explained in his keynote. 

Agent Kit is an integrated development environment for creating autonomous AI workers. It features a visual canvas to design complex workflows, an embeddable chat interface ("Chat Kit") for deploying agents in any app, and a sophisticated evaluation suite to measure and improve performance.

A compelling demo from financial operations platform Ramp showed how Agent Kit was used to build a procurement agent. An employee could simply type, "I need five more ChatGPT business seats," and the agent would parse the request, check it against company expense policies, find vendor details, and prepare a virtual credit card for the purchase — a process that once took weeks now completed in minutes. 

This push into agents is a direct response to a growing enterprise need to move beyond AI as a simple information retrieval tool and toward AI as a productivity engine that automates complex business processes. Brad Lightcap, OpenAI's COO, noted that for enterprise adoption, "you needed this kind of shift to more agentic AI that could actually do things for you, versus just respond with text outputs." 

The future of code and the Jony Ive bBombshell

Perhaps the most profound shift is occurring in software development itself. Codex, OpenAI's AI coding agent, has graduated from a research preview to a full-fledged product, now powered by a specialized version of the new GPT-5 model. It is, as one speaker put it, "a teammate that understands your context." 

The capabilities are staggering. Developers can now assign Codex tasks directly from Slack, and the agent can autonomously write code, create pull requests, and even review other engineers' work on GitHub. A live demo showed Codex taking a simple photo of a whiteboard sketch and turning it into a fully functional, beautifully designed mobile app screen. Another demo showed an app that could "self-evolve," reprogramming itself in real-time based on a user's natural language request. 

But the day's biggest surprise came in a closing fireside chat, which was not livestreamed, between Altman and Jony Ive, the iconic former chief design officer of Apple. The two revealed they have been collaborating for three years on a new family of AI-centric hardware.

Ive, whose design philosophy shaped the iPhone, iMac, and Apple Watch, said his creative team’s purpose "became clear" with the launch of ChatGPT. He argued that our current relationship with technology is broken and that AI presents an opportunity for a fundamental reset.

“I think it would be absurd to assume that you could have technology that is this breathtaking, delivered to us through legacy products, products that are decades old,” Ive said. “I see it as a chance to use this most remarkable capability to full-on address a lot of the overwhelm and despair that people feel right now.”

While details of the devices remain secret, Ive spoke of his motivation in deeply human terms. “We love our species, and we want to be useful. We think that humanity deserves much better than humanity generally is given,” he said. He emphasized the importance of "care" in the design process, stating, "We sense when people have cared... you sense carelessness. You sense when somebody does not care about you, they care about money and schedule." 

This collaboration confirms that OpenAI's ambitions are not confined to the cloud; it is actively exploring the physical interface through which humanity will interact with its powerful new intelligence.

The Unquenchable Thirst for Compute

Underpinning this entire platform strategy is a single, overwhelming constraint: the availability of computing power. In both the private press conference and the un-streamed Developer State of the Union, OpenAI’s leadership returned to this theme again and again.

“The degree to which we are all constrained by compute... Everyone is just so constrained on being able to offer the services at the scale required to get the revenue that at this point, we're quite confident we can push it pretty far,” Altman told reporters. He added that even with massive new hardware partnerships with AMD and others, "we'll be saying the same thing again. We're so convinced... There's so much more demand." 

This explains the company’s aggressive, multi-billion-dollar investment in infrastructure. When asked about profitability, Altman was candid that the company is in a phase of "investment and growth." He invoked a famous quote from Walt Disney, paraphrasing, "We make more money so we can make more movies." For OpenAI, the "movies" are ever-more-powerful AI models.

Greg Brockman, OpenAI’s President, put the ultimate goal in stark economic terms during the Developer State of the Union. "AI is going to become, probably in the not too distant future, the fundamental driver of economic growth," he said. "Asking ‘How much compute do you want?’ is a little bit like asking how much workforce do you want? The answer is, you can always get more out of more." 

As the day concluded and developers mingled at the reception, the scale of OpenAI's project came into focus. Fueled by new models like the powerful GPT-5 Pro and the stunning Sora 2 video generator, the company is no longer just building AI. It is building the world where AI will live — a world of intelligent apps, autonomous agents, and new physical devices, betting that in the near future, intelligence itself will be the ultimate platform.

Some of the largest providers of large language models (LLMs) have sought to move beyond multimodal chatbots — extending their models out into "agents" that can actually take more actions on behalf of the user across websites. Recall OpenAI's ChatGPT Agent (formerly known as "Operator") and Anthropic's Computer Use, both released over the last two years.

Now, Google is getting into that same game as well. Today, the search giant's DeepMind AI lab subsidiary unveiled a new, fine-tuned and custom-trained version of its powerful Gemini 2.5 Pro LLM known as "Gemini 2.5 Pro Computer Use," which can use a virtual browser to surf the web on your behalf, retrieve information, fill out forms, and even take actions on websites — all from a user's single text prompt.

"These are early days, but the model’s ability to interact with the web – like scrolling, filling forms + navigating dropdowns – is an important next step in building general-purpose agents," said Google CEO Sundar Pichai, as part of a longer statement on the social network, X.

The model is not available for consumers directly from Google, though.

Instead, Google partnered with another company, Browserbase, founded by former Twilio engineer Paul Klein in early 2024, which offers virtual "headless" web browser specifically for use by AI agents and applications. (A "headless" browser is one that doesn't require a graphical user interface, or GUI, to navigate the web, though in this case and others, Browserbase does show a graphical representation for the user).

Users can demo the new Gemini 2.5 Computer Use model directly on Browserbase here and even compare it side-by-side with the older, rival offerings from OpenAI and Anthropic in a new "Browser Arena" launched by the startup (though only one additional model can be selected alongside Gemini at a time).

For AI builders and developers, it's being made as a raw, albeit propreitary LLM through the Gemini API in Google AI Studio for rapid prototyping, and Google Cloud's Vertex AI model selector and applications building platform.

The new offering builds on the capabilities of Gemini 2.5 Pro, released back in March 2025 but which has been updated significantly several times since then, with a specific focus on enabling AI agents to perform direct interactions with user interfaces, including browsers and mobile applications.

Overall, it appears Gemini 2.5 Computer Use is designed to let developers create agents that can complete interface-driven tasks autonomously — such as clicking, typing, scrolling, filling out forms, and navigating behind login screens.

Rather than relying solely on APIs or structured inputs, this model allows AI systems to interact with software visually and functionally, much like a human would.

Brief User Hands-On Tests

In my brief, unscientific initial hands-on tests on the Browserbase website, Gemini 2.5 Computer Use successfully navigate to Taylor Swift's official website as instructed and provided me a summary of what was being sold or promoted at the top — a special edition of her newest album, "The Life of A Showgirl."

In another test, I asked Gemini 2.5 Computer Use to search Amazon for highly rated and well-reviewed solar lights I could stake into my back yard, and I was delighted to watch as it successfully completed a Google Search Captcha designed to weed out non-human users ("Select all the boxes with a motorcycle.") It did so in a matter of seconds.

However, once it got through there, it stalled and was unable to complete the task, despite serving up a "task competed" message.

I should also note here that while the ChatGPT agent from OpenAI and Anthropic's Claude can create and edit local files — such as PowerPoint presentations, spreadsheets, or text documents — on the user’s behalf, Gemini 2.5 Computer Use does not currently offer direct file system access or native file creation capabilities.

Instead, it is designed to control and navigate web and mobile user interfaces through actions like clicking, typing, and scrolling. Its output is limited to suggested UI actions or chatbot-style text responses; any structured output like a document or file must be handled separately by the developer, often through custom code or third-party integrations.

Performance Benchmarks

Google says Gemini 2.5 Computer Use has demonstrated leading results in multiple interface control benchmarks, particularly when compared to other major AI systems including Claude Sonnet and OpenAI’s agent-based models.

Evaluations were conducted via Browserbase and Google’s own testing.

Some highlights include:

• Online-Mind2Web (Browserbase): 65.7% for Gemini 2.5 vs. 61.0% (Claude Sonnet 4) and 44.3% (OpenAI Agent)

• WebVoyager (Browserbase): 79.9% for Gemini 2.5 vs. 69.4% (Claude Sonnet 4) and 61.0% (OpenAI Agent)

• AndroidWorld (DeepMind): 69.7% for Gemini 2.5 vs. 62.1% (Claude Sonnet 4); OpenAI's model could not be measured due to lack of access

• OSWorld: Currently not supported by Gemini 2.5; top competitor result was 61.4%

In addition to strong accuracy, Google reports that the model operates at lower latency than other browser control solutions — a key factor in production use cases like UI automation and testing.

How It Works

Agents powered by the Computer Use model operate within an interaction loop. They receive:

• A user task prompt

• A screenshot of the interface

• A history of past actions

The model analyzes this input and produces a recommended UI action, such as clicking a button or typing into a field.

If needed, it can request confirmation from the end user for riskier tasks, such as making a purchase.

Once the action is executed, the interface state is updated and a new screenshot is sent back to the model. The loop continues until the task is completed or halted due to an error or a safety decision.

The model uses a specialized tool called computer_use, and it can be integrated into custom environments using tools like Playwright or via the Browserbase demo sandbox.

Use Cases and Adoption

According to Google, teams internally and externally have already started using the model across several domains:

• Google’s payments platform team reports that Gemini 2.5 Computer Use successfully recovers over 60% of failed test executions, reducing a major source of engineering inefficiencies.

• Autotab, a third-party AI agent platform, said the model outperformed others on complex data parsing tasks, boosting performance by up to 18% in their hardest evaluations.

• Poke.com, a proactive AI assistant provider, noted that the Gemini model often operates 50% faster than competing solutions during interface interactions.

The model is also being used in Google’s own product development efforts, including in Project Mariner, the Firebase Testing Agent, and AI Mode in Search.

Safety Measures

Because this model directly controls software interfaces, Google emphasizes a multi-layered approach to safety:

• A per-step safety service inspects every proposed action before execution.

• Developers can define system-level instructions to block or require confirmation for specific actions.

• The model includes built-in safeguards to avoid actions that might compromise security or violate Google’s prohibited use policies.

For example, if the model encounters a CAPTCHA, it will generate an action to click the checkbox but flag it as requiring user confirmation, ensuring the system does not proceed without human oversight.

Technical Capabilities

The model supports a wide array of built-in UI actions such as:

• click_at, type_text_at, scroll_document, drag_and_drop, and more

• User-defined functions can be added to extend its reach to mobile or custom environments

• Screen coordinates are normalized (0–1000 scale) and translated back to pixel dimensions during execution

It accepts image and text input and outputs text responses or function calls to perform tasks. The recommended screen resolution for optimal results is 1440x900, though it can work with other sizes.

API Pricing Remains Almost Identical to Gemini 2.5 Pro

The pricing for Gemini 2.5 Computer Use aligns closely with the standard Gemini 2.5 Pro model. Both follow the same per-token billing structure: input tokens are priced at $1.25 per one million tokens for prompts under 200,000 tokens, and $2.50 per million tokens for prompts longer than that.

Output tokens follow a similar split, priced at $10.00 per million for smaller responses and $15.00 for larger ones.

Where the models diverge is in availability and additional features.

Gemini 2.5 Pro includes a free tier that allows developers to use the model at no cost, with no explicit token cap published, though usage may be subject to rate limits or quota constraints depending on the platform (e.g. Google AI Studio).

This free access includes both input and output tokens. Once developers exceed their allotted quota or switch to the paid tier, standard per-token pricing applies.

In contrast, Gemini 2.5 Computer Use is available exclusively through the paid tier. There is no free access currently offered for this model, and all usage incurs token-based charges from the outset.

Feature-wise, Gemini 2.5 Pro supports optional capabilities like context caching (starting at $0.31 per million tokens) and grounding with Google Search (free for up to 1,500 requests per day, then $35 per 1,000 additional requests). These are not available for Computer Use at this time.

Another distinction is in data handling: output from the Computer Use model is not used to improve Google products in the paid tier, while free-tier usage of Gemini 2.5 Pro contributes to model improvement unless explicitly opted out.

Overall, developers can expect similar token-based costs across both models, but they should consider tier access, included capabilities, and data use policies when deciding which model fits their needs.

For more than a decade, conversational AI has promised human-like assistants that can do more than chat. Yet even as large language models (LLMs) like ChatGPT, Gemini, and Claude learn to reason, explain, and code, one critical category of interaction remains largely unsolved — reliably completing tasks for people outside of chat.

Even the best AI models score only in the 30th percentile on Terminal-Bench Hard, a third-party benchmark designed to evaluate the performance of AI agents on completing a variety of browser-based tasks, far below the reliability demanded by most enterprises and users. And task-specific benchmarks like TAU-Bench airline, which measures the reliability of AI agents on finding and booking flights on behalf of a user, also don't have much higher pass rates, with only 56% for the top performing agents and models (Claude 3.7 Sonnet) — meaning the agent fails nearly half the time.

New York City-based Augmented Intelligence (AUI) Inc., co-founded by Ohad Elhelo and Ori Cohen, believes it has finally come with a solution to boost AI agent reliability to a level where most enterprises can trust they will do as instructed, reliably.

The company’s new foundation model, called Apollo-1 — which remains in preview with early testers now but is close to an impending general release — is built on a principle it calls stateful neuro-symbolic reasoning.

It's a hybrid architecture championed by even LLM skeptics like Gary Marcus, designed to guarantee consistent, policy-compliant outcomes in every customer interaction.

“Conversational AI is essentially two halves,” said Elhelo in a recent interview with VentureBeat. “The first half — open-ended dialogue — is handled beautifully by LLMs. They’re designed for creative or exploratory use cases. The other half is task-oriented dialogue, where there’s always a specific goal behind the conversation. That half has remained unsolved because it requires certainty.”

AUI defines certainty as the difference between an agent that “probably” performs a task and one that almost “always” does.

For example, on TAU-Bench Airline, it performs at a staggering 92.5% pass rate, leaving all the other current competitors far behind in the dust — according to benchmarks shared with VentureBeat and posted on AUI's website.

Elhelo offered simple examples: a bank that must enforce ID verification for refunds over $200, or an airline that must always offer a business-class upgrade before economy.

“Those aren’t preferences,” he said. “They’re requirements. And no purely generative approach can deliver that kind of behavioral certainty.”

AUI and its work on improving reliability was previously covered by subscription news outlet The Information, but has not received widespread coverage in publicly accessible media — until now.

From Pattern Matching to Predictable Action

The team argues that transformer models, by design, can’t meet that bar. Large language models generate plausible text, not guaranteed behavior. “When you tell an LLM to always offer insurance before payment, it might — usually,” Elhelo said. “Configure Apollo-1 with that rule, and it will — every time.”

That distinction, he said, stems from the architecture itself. Transformers predict the next token in a sequence. Apollo-1, by contrast, predicts the next action in a conversation, operating on what AUI calls a typed symbolic state.

Cohen explained the idea in more technical terms. “Neuro-symbolic means we’re merging the two dominant paradigms,” he said. “The symbolic layer gives you structure — it knows what an intent, an entity, and a parameter are — while the neural layer gives you language fluency. The neuro-symbolic reasoner sits between them. It’s a different kind of brain for dialogue.”

Where transformers treat every output as text generation, Apollo-1 runs a closed reasoning loop: an encoder translates natural language into a symbolic state, a state machine maintains that state, a decision engine determines the next action, a planner executes it, and a decoder turns the result back into language. “The process is iterative,” Cohen said. “It loops until the task is done. That’s how you get determinism instead of probability.”

A Foundation Model for Task Execution

Unlike traditional chatbots or bespoke automation systems, Apollo-1 is meant to serve as a foundation model for task-oriented dialogue — a single, domain-agnostic system that can be configured for banking, travel, retail, or insurance through what AUI calls a System Prompt.

“The System Prompt isn’t a configuration file,” Elhelo said. “It’s a behavioral contract. You define exactly how your agent must behave in situations of interest, and Apollo-1 guarantees those behaviors will execute.”

Organizations can use the prompt to encode symbolic slots — intents, parameters, and policies — as well as tool boundaries and state-dependent rules.

A food delivery app, for example, might enforce “if allergy mentioned, always inform the restaurant,” while a telecom provider might define “after three failed payment attempts, suspend service.” In both cases, the behavior executes deterministically, not statistically.

Eight Years in the Making

AUI’s path to Apollo-1 began in 2017, when the team started encoding millions of real task-oriented conversations handled by a 60,000-person human agent workforce.

That work led to a symbolic language capable of separating procedural knowledge — steps, constraints, and flows — from descriptive knowledge like entities and attributes.

“The insight was that task-oriented dialogue has universal procedural patterns,” said Elhelo. “Food delivery, claims processing, and order management all share similar structures. Once you model that explicitly, you can compute over it deterministically.”

From there, the company built the neuro-symbolic reasoner — a system that uses the symbolic state to decide what happens next rather than guessing through token prediction.

Benchmarks suggest the architecture makes a measurable difference.

In AUI’s own evaluations, Apollo-1 achieved over 90 percent task completion on the τ-Bench-Airline benchmark, compared with 60 percent for Claude-4.

It completed 83 percent of live booking chats on Google Flights versus 22 percent for Gemini 2.5-Flash, and 91 percent of retail scenarios on Amazon versus 17 percent for Rufus.

“These aren’t incremental improvements,” said Cohen. “They’re order-of-magnitude reliability differences.”

A Complement, Not a Competitor

AUI isn’t pitching Apollo-1 as a replacement for large language models, but as their necessary counterpart. In Elhelo’s words: “Transformers optimize for creative probability. Apollo-1 optimizes for behavioral certainty. Together, they form the complete spectrum of conversational AI.”

The model is already running in limited pilots with undisclosed Fortune 500 companies across sectors including finance, travel, and retail.

AUI has also confirmed a strategic partnership with Google and plans for general availability in November 2025, when it will open APIs, release full documentation, and add voice and image capabilities. Interested potential customers and partners can sign up to receive more information when it becomes available on AUI's website form.

Until then, the company is keeping details under wraps. When asked about what comes next, Elhelo smiled. “Let’s just say we’re preparing an announcement,” he said. “Soon.”

Toward Conversations That Act

For all its technical sophistication, Apollo-1’s pitch is simple: make AI that businesses can trust to act — not just talk. “We’re on a mission to democratize access to AI that works,” Cohen said near the end of the interview.

Whether Apollo-1 becomes the new standard for task-oriented dialogue remains to be seen. But if AUI’s architecture performs as promised, the long-standing divide between chatbots that sound human and agents that reliably do human work may finally start to close.

Otter is launching a suite of enterprise tools to help companies use its tech to create a central knowledge base.

As AI drives unprecedented demand for data centers, the UK must pioneer biodiversity-positive development, writes David Davies.

Major alternative asset management firms are heavily investing in data center projects as a way to capitalize on AI-driven demand for processing power.

As the data center industry tackles soaring costs, Alaska could cut cooling bills by as much as $150 million a year for a 1 GW campus, according to Governor Mike Dunleavy.