Menu Close
Scale AI

Scale AI

Scale AI News: AI Benchmark Reveals Knowledge Gaps

January 23, 2025

Scale AI Unicorn News - January 23, 2025

Scale AI and the Center for AI Safety have released 'Humanity’s Last Exam,' a benchmark testing AI models against expert-level questions. The results show current AI systems answering fewer than 10% correctly, highlighting significant gaps in AI reasoning and knowledge. The dataset will be shared with the research community to drive further advancements.

( 00:00:00 ) Introduction

( 00:00:28 ) Scale AI Unveils Results of Humanity's Last Exam

play
0:00 - 0:00
play button

Scale AI has introduced the SEAL (Scale Expert AI Leaderboards) initiative, featuring private benchmarks designed to provide unbiased evaluations of leading AI models. These benchmarks are continuously updated and conducted by domain experts to ensure accuracy and relevance. Additionally, in partnership with the Center for AI Safety, Scale AI has launched "Humanity's Last Exam," a project aimed at developing the world's most challenging AI benchmark to assess expert-level AI capabilities. These efforts highlight the need for more sophisticated evaluations as AI systems rapidly advance, often surpassing traditional benchmarks.

Empower Your Portfolio with Private Equity

Invest in Scale AI

Disclaimer

Investing in private securities is speculative, illiquid, and involves risk of loss. An investment with Linqto is a private placement and does not grant or transfer ownership of private company stock. No guarantee is made that a company will experience an IPO or any liquidity event.

Linqto leverages advanced artificial intelligence (AI) technologies to generate Unicorn News to summarize updates about private companies. The news summaries and audio are both AI generated, based on the source(s) listed.