The Epoch Brief - June 26, 2026
Our new long-horizon coding benchmark, hyperscaler cash depletion, tracking AI R&D automation, and Chinese lab strategies
Welcome to another edition of the Epoch Brief! Here’s what’s new:
The launch of MirrorCode, our new long-horizon coding benchmark, co-developed with METR, to better measure the limits of autonomous AI coding abilities.
Hyperscaler capex will overtake operating cash flows by the end of 2026.
Two new Gradient Updates: what is learned from analyzing 1,604 job postings by Chinese AI labs, and a proposed AI R&D taxonomy to track which parts of research remain unautomated.
We're hiring two designers to turn complex research into dashboards and visualizations that researchers and policymakers can easily use.
MirrorCode: A benchmark for real-world software engineering
MirrorCode, co-developed with METR, is our new long-horizon benchmark to answer the question: What's the largest software project AI can complete on its own? The benchmark tasks AI models with rebuilding 25 real-world programs spanning bioinformatics, Unix utilities, cryptography, interpreters, and more. No access to source code and no human in the loop. We estimate that the hardest programs would take a human engineer, without AI assistance, weeks to months to complete.
MirrorCode provides AI models with a large enough inference budget to make a serious attempt at real-world software engineering (SWE) tasks. Many existing SWE benchmarks cap inference at around $1-$10 per task, with runs lasting only minutes, or, at most, hours. By comparison, one of the largest MirrorCode tasks cost $2,600 for a single run and involved AI working for 19 days without human intervention.
Claude Opus 4.7 leads other models with a 56% solve rate, meaning there is significant room for further improvement. View the full results and analysis.
Data Insight: Hyperscaler capex is on trend to outpace their cash inflows by the end of 2026
Senior researcher Isabel Juniewicz finds that the world’s largest hyperscalers (Microsoft, Amazon, Alphabet, Meta, and Oracle) are increasing their cash capital expenditures faster than their cash inflows from operations. Most hyperscalers have already turned to external financing to fund their growing investments in AI infrastructure, or are considering doing so.
Commentary
We published two new Gradient Updates, where Epoch researchers and guests share more opinionated or informal takes on big questions in AI progress. Gradient Updates represent the views of the authors, and do not necessarily reflect the views of Epoch AI as a whole.
What we learned from 1,604 Chinese AI job postings
Cheryl Wu, JS Denain, and Anson Ho scraped more than 1600 job postings from six major Chinese firms to better understand their strategies. They find that, like US firms, Chinese labs aren’t all following the same playbook and have distinct “personalities”.
Toward an O*NET for AI R&D
How close is AI to automating AI research and development? Right now, the tools economists use to track automation are too blunt to say. Joe Kwon, threat modeling and catastrophic risk reduction researcher, alongside Epoch’s Jean-Stanislas Denain and Anson Ho, propose a sharper tool: a thorough taxonomy of 60+ tasks involved in frontier AI research.
Other Updates
Careers
We’re hiring across several roles. All positions are fully remote.
Senior Product Designer and Product Designer to translate complex research into intuitive, engaging, and high-impact designs.
Researchers and Senior Researchers to lead new projects across our expanding teams.
Data Scientist (Contract) to assist with our AI research efforts, including reviewing technical literature, tracking benchmark data, and analyzing AI models, data centers, and companies.
Applications are rolling, so apply soon!
Asking LLMs to reimplement existing software libraries while giving the input-output access to the original program, documentation and test examples is really genius level stuff. And if we look at the libraries being implemented they are essentially collections of standard standalone functions, for which there are droves of examples in the training data, with little architectural complexity. In other words most of the tasks in the benchmark reduce to implementing a large number of individual functions that are straightforward to implement using vanilla data structures i.e. can be broken down to bite-sized subtasks that are just right for LLMs to succeed on. And despite this ideal setup (and despite throwing up to $2,600 and 19 days of compute at a single task) the grand result is that the absolute best model can only reliably implement 56% of the programs, a number heavily skewed towards the simplest tasks.