PaperBench: Evaluating AI’s Ability to Replicate AI Research
We introduce PaperBench, a benchmark evaluating the ability of AI agents to replicate state-of-the-art AI research.
Story
OpenAI · Artificial Intelligence
Join the conversation
Discuss the story with the TraceFeed community.
No comments yet. Be the first to add a perspective.