Artificial Analysis Publishes Full Eval Breakdown for Claude Sonnet 5.5 Across Reasoning Efforts

ArtificialAnlys · x · 2026-09-29

Artificial Analysis released the full breakdown of individual evaluations in its Intelligence Index for Claude Sonnet 5.5 across all reasoning effort levels, supplementing its finding that max effort burns 193k output tokens per task.

Related event: Sonnet 5.5 Nearly Matches Opus 5.5 but Sets Token Consumption Record(12 posts)→

Original post →

More from Models

Models channel →