Induction benchmark: Astra first, GPT-5.6 Sol third, GPT-6 Sol results pending

Serafim Batzoglou @s_batzoglouOriginally posted 2026.09.23

Open original on X

Opus 5.5 is an excellent model! I benchmarked it on induction, and it comes a clear second to Astra, with a distant third being GPT 5.6 Sol. GPT 6 Sol and Luna are taking their time returning results, will be updating at some point. Github: https://github.com/SerafimBatzoglou/concept-synth/blob/main/benchmarks/induction/docs/leaderboard.md Paper (ICML spotlight 2026): https://arxiv.org/abs/2602.18956