Reaction
Sonnet 5.5 tops the Artificial Analysis Coding Agent Index at max effort; xhigh scores 63 versus 68
Artificial Analysis reports per-task API costs of $3.33 at xhigh and $14.19 at max for Sonnet 5.5.
TLDR
Artificial Analysis says Sonnet 5.5 in Claude Code scores 68 at max effort and 63 at xhigh on its Coding Agent Index. It puts per-task API cost at $14.19 versus $3.33, a 77% reduction at xhigh. Max generated 4.8 times as many output tokens and took 3.4 times as many turns; its largest benchmark gain was on Terminal-Bench 4.0, rising from 58.1% to 66.2%.
Combined views
—
3 Sources, first seen ago
