STORY · FORSKNING_

How large software projects can AI complete on its own?

The MirrorCode benchmark gives AI sufficient compute budget to attempt larger software tasks. Some tasks cost $2,600 and let AI work for 19 days without human intervention, in contrast to traditional benchmarks that limit budgets to $1–10.

WHY IT MATTERS

This reveals the scale of tasks AI can now handle alone, and how realistic evaluation frameworks uncover greater capabilities than previous benchmarks.

SOURCES

MACHINE-GENERATED SUMMARY This summary is written by machine from the sources below. We sort and explain — but we are a way into the field, not the final word. Check the source when something matters to you.