Скачать или смотреть LiveCodeBench Pro: AI for Competitive Programming | Zerui Cheng | Princeton University

At the Open AGI Symposium at the University of California, Berkeley on August 4th, we were joined by Zerui Cheng of Princeton University, who introduced LiveCodeBench Pro, the future gold-standard benchmark for testing AI on competitive programming problems. He explained how it measures deep algorithmic reasoning in large language models and why it's already being used by OpenAI, Google, and Anthropic.

Agenda:

Motivation – What is Competitive Programming (CP) and why it matters

Current Gaps – Limitations of current CP benchmarks

Our Solution and Main Results – A quick glance at LiveCodeBench Pro

Deep Diagnostics – Fine-grained annotations and error analysis

Open Questions and Discussion

Q&A

Learn more here - https://livecodebenchpro.com/

Stay Updated Here:

Twitter - https://x.com/openagisummit
Website - https://openagi.xyz/