Run Characteristics Across Models and Effort Levels
Bubble chart of run characteristics: each bubble represents one benchmark run, positioned by runtime and chronology event output, sized by tokens consumed.
Model (fill) Haiku 4.5 Sonnet 4.6 Opus 4.7
Effort (border)
Low Med High Xhigh Max
Bubble area ∝ tokens consumed (85k–249k).