Vercel AI Gateway Index: Jev Reached 35.3% of Paid Teams in 14 Days for 0.27% of Spend
Vercel’s October AI Gateway Index crowned Jev the fastest-adopted model in the platform’s history—nearly double Claude Sonnet 5.5’s reach while consuming a tiny fraction of spend.
Concrete production telemetry backed up Jev’s viral launch buzz on October 16 when Vercel published its October 2026 AI Gateway Index, revealing that Jev was adopted faster than any other model Vercel has ever measured.
Record reach at 0.27% of total spend
Within its first 14 days on Vercel’s AI Gateway, Jev was actively used by 35.3% of paid teams—nearly double the 18.1% two-week adoption rate of second-place Claude Sonnet 5.5. Even more striking was the billing share: despite massive request volume across more than a third of paid organizations, Jev accounted for just 0.27% of gateway spend thanks to its $0.042 per million input token pricing and free outputs.
Vercel itself became an internal production user, swapping Jev in place of GPT-6 Luna to classify harmful commands at lower latency and cost.
The Jevons paradox in action
Combined with later figures showing decision models accounting for one in four requests across Vercel’s gateway during October, the report illustrated the economic dynamic behind Jev’s name. When semantic judgment drops two orders of magnitude in price and latency, teams stop rationing AI calls to top-level chat boxes and begin sprinkling them throughout middleware, guardrails, and UI interactions.