Hacker News Reader

okay what's a point in time where Claude was better? just give me a date

until GPT 5.4, Claude always had a decent edge in benchmarks (any date before then). The gap was huge during the Sonnet 3.x vs GPT 4.x eta.