Vercel data: open models now handle 71.5% of its AI gateway tokens

Vercel's October AI Gateway Production Index, based on anonymised traffic through its AI gateway up to September, says open-weight models processed 71.5% of tokens and about 23% of spend, up from one in twenty dollars in June. Gateway spend in September was about 13 times December's while the average price per token fell 47%, and almost one in three teams changed their most-used AI lab in September. DeepSeek-V4.1-Flash alone handled nearly half of September's tokens for 2.6% of spend. Jev, which Vercel describes as a decision model from TypeSafe AI, reached 35% of paid teams in its first two weeks, the fastest adoption Vercel has recorded, while taking only 0.27% of spend. This is one gateway's view, and spend is estimated at list prices, so actual bills may differ.