Sakana AI's Agent Beats 800+ Humans to Win a Competitive Programming Contest

Sakana AI's ALE-Agent took first place in an AtCoder Heuristic Contest, autonomously discovering novel optimization algorithms over a four-hour window — spending about $1,300 in API credits to do it.

An AI agent built by Sakana AI placed first in an AtCoder Heuristic Contest, beating over 800 human competitors in a production planning optimization challenge, as @SakanaAILabs announced on Monday. The contest ran for four hours, during which the agent — called ALE-Agent — autonomously iterated on heuristic algorithms using inference-time scaling with frontier language models. It wasn't simply applying known techniques; Sakana claims the agent invented novel heuristics to solve the problem.

Unlock the full briefing

Get every story in today's briefing, the full archive, and the daily AI intelligence brief.

All stories today

Full archive

Daily brief

Cancel anytime. Payments powered by Stripe.