Anthropic's Claude Opus 4.5 hits nearly five hours of autonomous agentic work, signaling a super-exponential AI capability curve
Solo commentary by Nate B Jones of AI News & Strategy Daily on what a new agentic benchmark result means for the future of work.
Summary
Nate B Jones presents the case that AI development has entered a super-exponential growth phase, using a new benchmark result from Anthropic's Claude Opus 4.5 as the central evidence. The model achieved nearly five hours of sustained, useful agentic work at a 50% success rate — a dramatic leap from the minutes-long horizons measured not long ago. Jones argues that the doubling rate of approximately every four to four-and-a-half months means that by end of 2026, AI agents may be capable of performing a full week's worth of human work autonomously. He contends this is not hype but a mathematical reality that will restructure careers, job families, and the distribution of economic value — and that people who begin learning to assign and manage agent work now will have a compounding advantage over those who wait.
Key Takeaways
FULL TRANSCRIPT
The METR Benchmark and Why It Matters
Nate B Jones: We are on a super-exponential timeline for AI agents, and I want to explain what that means and why it's critically important that we all pay attention to it.
METR is the Model Evaluation and Threat Research company. It's a nonprofit dedicated to understanding how models perform, and they are famous for producing a graph that shows how long models can do useful agentic work at a time. It's a somewhat confusing graph to understand, so I'm going to explain it simply.
Basically, they take a task and measure how long a human takes to do that work task. Then they want to find out if the AI can do that task with at least a 50% likelihood of success. Why 50%? Because they had to pick a number somewhere. They also measure it at 80%, and we'll get to that.
METR is important because it does not top out. If you have benchmarks like SWE-bench — which is an engineering one — it tops out at 100%, and we're already way up at the top. You can go from 91% to 93% and you don't really get a sense of how the models are changing. METR is different because that graph has no top end. It can just keep measuring more and more work, and that allows it to show super-exponential progress.
The Super-Exponential Debate and the Opus 4.5 Result
One of the biggest debates of 2025 was: are we on an exponential timescale with AI, or are we on a super-exponential — where progress is increasing faster than exponentially? It seems like we are on the super-exponential trend line. One of the things that made us think that is this latest result from Claude Opus 4.5, which shows over four hours — four hours and 45 minutes, almost five hours — of human-equivalent work done at a 50% likelihood of success.
The 80% mark is also measured, and it is 27 to 28 minutes for Opus 4.5. You might think that's not that far, but keep in mind it was not that long ago that we were at one minute, two minutes, ten minutes, thirty minutes — and now we're up to almost five hours. That is the point of a super-exponential curve.
We are on a doubling rate of every four to four-and-a-half months right now. So if the number is 50% complete but the time horizon is nearly five hours, we're going to be at ten hours by the end of Q1. We'll be at twenty hours by the end of Q2 into Q3. We may be at forty hours by the end of the year or beyond. And that is why we have to pay attention to this.
The Self-Reinforcing Flywheel
Super-exponential gains suggest that we have hit a self-reinforcing flywheel with AI. That is indeed what we hear from model makers, and that is why 2025 was the last normal year. We are going to see really remarkable progress from AI in 2026 and every year after, because AI itself is starting to reinforce AI systems. We're bringing AI in to help train AI systems, and that is going to become more and more automated. We are going to have capabilities that AI itself helps to grow, speeding up the whole process — and all of that is going to allow us to continue making progress on these tough tasks that have no upper limit.
What This Means for Work and Careers
This matters because our ability to do meaningful work is going to be determined by whether or not we can define useful, high-taste, high-quality work that an AI can do over a period of time. Do you have something for an AI that would take you a week to do? Maybe it's your taxes. But that is going to increasingly become the question. And if you don't, then the question becomes: what does it take for you to get there? What does it take for you to gain the skill to assign that work?
Because in a super-exponential world, the skill we need to learn is also super-exponential. The people who figure out how to assign agents work now — in January, February, and March — are going to have a much easier time learning how to continue assigning agents work when the agents can do much harder things. Whereas if you wait and say, "I'm going to catch up — I've scheduled this for Q2 or Q3 next year, that's my AI quarter" — good luck with that. It doesn't work that way. There will be people running circles around you because they can assign their agents a week's worth of work. And once you can assign your agents a week's worth of work and spin up two or three of them, look at how much more productive that makes you. You're going to be running circles around people. That is the power-law distribution world we're going to live in.
Super-Exponentials Create Power Laws
Super-exponentials create power laws. A power law is the idea that the world we live in is not normally distributed. In a normally distributed world, most people cluster around the average, with a few people on the tails — Einstein is way out here on one end. But in a power-law world, just a few people are going to be able to do a tremendous amount. And it's not because they're necessarily going to have lots of money to do it. It's because they have the skills to do it.
AI is going to disproportionately reward skill development where it's related to artificial intelligence, and in everything else, people are going to start to lose traction. If you are looking to make a dent in your career, I would look less in 2026 at your job family's traditional requirements, and look more at: where can an agent do a meaningful amount of work for a week in this traditional job-family area, and how can I make sure I set myself up so I know how to define and assign that work, know how to hold it accountable, know how to put good taste down so I know what excellent looks like in that work, know how to intervene, keep the agent on track, and have the technical foundations necessary to define and set up an agentic system.
This is going to become more and more relevant for all of us. The technical skill sets are going to spread across job families. The non-technical skill sets are also going to spread across job families. Engineers who traditionally just had to write code are going to have to have some business fluency and customer fluency now, because they have to be the ones with good taste when they're architecting systems. And frankly, they now have to architect systems that non-technical people can contribute code to.
So just that one shift — the ability of agents to do work over time — is going to multiply the impacts across all of the rest of us. Having agents that work longer means all of our jobs are going to change forever.
The Reality of the Curve
You might think I'm a hype person. This is not me being hyped up. This is me talking about the reality that we are on a super-exponential curve. Humans are bad at estimating super-exponential curves, and so I just want to make it really concrete. There is no way that work will not change for everybody if we are in a place where it's five hours and doubling every four months. By April you're going to be at ten hours. By July or September you're going to be at twenty hours. By December you're going to be at forty hours — maybe. It's just going to be extraordinary.
Are you able to delegate a week's worth of work? That is the question of 2026.
We will all have to let go of a lot. We will have to let go of our traditional understandings about career progression. We'll have to let go of our traditional understandings about job families — what job families know and what they don't. We are going to have to be outcome-obsessed and ownership-obsessed. The work of the future is going to reward people who are ownership- and outcome-obsessed, because that's where human value shows up. It's when we make sure that what's made is actually relevant for people, actually useful, actually good. It's not just vibe-coded slop.
There will be lots of vibe-coded slop. In fact, I would expect it to 100x in 2026, because you can ask your agent to do a lot of terrible things. It's going to be up to you to decide that the agent's work is worth it — that you are assigning the agent meaningful work and the agent is doing good work that compounds over time.
Strategy Becomes an Individual Skill
The strategic rewards used to accrue to leaders. Strategy is now an individual thing, because you are effectively a strategic manager of a team of agents — or you will be in 2026. You can make them yourself. There will probably be startups that market them to you. But either way, you will end up with a team of agents working for you. Do you know how to manage them? Do you know how to lead them? Do you know how to drive them to develop compounding advantage over time? That used to be a question for directors and above. It's not anymore — it's for everybody.
Everyone will need to be able to do this, and the people who can are going to look like they can do anything. The span is going to be incredible, because they're able to leverage their own domain expertise and expand their scope of impact from there.
Domain Expertise Still Matters
I do not mean that you can do anything that requires deep domain expertise that you do not have. There are still going to be real values you can't get to just by adding agents. For example, if you are a lawyer with decades of experience, agents are going to transform the legal profession and how you work — but it's not going to transform it to the point where I, as a non-lawyer, can come in and do work for a white-shoe law firm and get exactly the same quality of work done at the end as the lawyer who has decades of experience. There is going to be a reward for understanding the business deeply, and that will show up in your ability to direct AI agents toward useful ends.
So as much as it may seem like I'm saying the agent can do the work and we won't do any — what I'm really saying is our domain expertise is worth more and more. But we have to be smart and leverage it really differently to get where we need to go in 2026. And that's going to change all of our skill sets. We're all going to have to learn together. We've never gone through this kind of workflow and workforce transformation before, so we're all going to have to jump in and figure out how to do it together.
The Road Ahead
I do think it's real. I do think it's coming. And I do think the key is that super-exponential graph. Opus 4.5 was just the latest model to reach five hours. It won't be the last. Claude doesn't have a special monopoly on this — we're going to see this from Gemini, we're going to see this from ChatGPT, we'll see it from other model makers as well. We will continue to see exponential gains in agent working time in 2026, and that will change the way all of us have to do our work.