The UK's AI Security Institute (AISI) tested frontier models across seven benchmarks with varying compute budgets. The finding: fixed budget caps systematically underestimate how capable AI agents really are.

An AI agent's performance is a curve that rises with test-time compute, the amount of processing power an agent is allowed to burn while working on a task. Cut the budget while the curve is still climbing, and the measured score tells you the minimum, not the maximum.

That's what the AISI researchers set out to prove in their latest work. The big question: how much do capabilities scale with compute, and what does that mean for cybersecurity?

More compute, more AI performance. But where's the limit? | Image: AISI

More compute, better results across the board