Benchmarks

Rank.ai’s blended leaderboard puts Claude Fable 5 narrowly ahead of GPT-5.6 Sol

An October 7 blended leaderboard gives Claude Fable 5 an 85–84 edge over GPT-5.6 Sol, but its component boards use different methods and update dates.