Grok Breaks 50% Barrier on EEbench, Ending Anthropic's Dominance

scaling01 · x · 2026-08-14

xAI's Grok model has officially taken the #2 spot on the EEbench electrical-engineering benchmark, breaking Anthropic's hold on the top of the leaderboard.

This milestone makes xAI only the second AI lab to cross the 50% mark on the benchmark. The achievement highlights that hardware engineering agents are rapidly becoming highly capable and practical.

Original post →

More from Models

Models channel →