Open-weight models surge to 29% of volume, price per token flattens
Vercel's AI Gateway Production Index for July 2026 reports that open-weight models have surged to 29% of total inference volume, up from previous periods. This shift reflects growing adoption of models with publicly available weights. Concurrently, the price per token has flattened, suggesting that the rapid cost declines seen earlier in the year have stabilized. The data is drawn from Vercel's AI Gateway, which proxies and monitors AI inference traffic for production applications. The index provides a snapshot of trends in model usage and pricing, highlighting the increasing share of open-weight models and a plateau in per-token costs.
// why it matters
Developers should note the rising adoption of open-weight models and stable pricing for inference costs.