Fable 5.2 benchmarks slightly better than Opus 5.5 - and my experience with it (which is limited due to how quickly it burns through tokens) suggests the same. If things were accelerating recursively (and results improving), wouldn't we be seeing increased release cadence and much greater improvement in capability?
Fable 5.2 benchmarks slightly better than Opus 5.5 - and my experience with it (which is limited due to how quickly it burns through tokens) suggests the same. If things were accelerating recursively (and results improving), wouldn't we be seeing increased release cadence and much greater improvement in capability?