Update on our long-horizon AI R&D evaluations

In April, we launched CRUX, a project to regularly run open-world evaluations: long, messy, real-world tests of what AI agents can actually do. Our second evaluation is underway, and we are asking: can AI agents automate AI research?

There is a lot of interest in studying AI research automation. But most of the systems built so far follow one of three patterns:

1. Keep a human in the loop to guide the agent and course-correct along the way.
2. Focus on narrow problems where ground truth is clear and progress is easy to verify, as in AutoResearch.
3. Use scaffolds engineered for one specific type of research question, so strong results may say more about the scaffold than about the agent's general research ability.

These efforts are helpful, but a lot of AI research is much broader. Success is not immediately clear or verifiable. Researchers need to test and reject promising hypotheses, backtrack, consider new or unconventional approaches, and do much more to make progress on answering research questions.

In CRUX 2, we are trying to test whether agents can answer novel, open-ended AI research questions.

One major risk in such a task is contamination. We want the agent to have access to the internet and all the tools it needs to solve the task, so we cannot use research questions from publicly available papers. At the same time, we want high-quality papers to serve as the source of challenging research questions.

To address this, we partnered with AI researchers from UK AISI, the University of Toronto, Princeton, and other institutions who have written high-quality papers that are not yet public, so there is no risk of contamination.

The authors pose open-ended research questions without giving away answers. The agent must produce a NeurIPS-quality paper and a reproducible codebase, which the authors of the papers then review.

We built a general-purpose scaffold on OpenClaw and Opus 4.8. We would have loved to use Fable 5, but given the filters on AI R&D capabilities, we do not want to confound the results.

Agents get generous resource budgets set in consultation with the original authors, such as access to VMs, GPUs, and any other compute needed to answer the question. They also have $3,000 in API credits per paper. We evaluate them on week-long time horizons to make progress on answering the research question, far more than typical agent evaluations.

The agent needs to manage its own budget. It can track its spend and stay within its limits, and it can modify its scaffold and reasoning effort as it sees fit.

In addition to the final artifacts, such as the paper and code, we are also evaluating the agent's trajectories in depth.

When we announced CRUX, we planned to conduct an open-world evaluation every month. Given the scope and ambition of this project, we have spent a lot more time making sure we are confident in our setup and results. That said, the early results we have are exciting, and we look forward to sharing them soon.