AI Gateway adds unified fast mode support
By Steven Van ·
AI Gateway has a unified fast mode abstraction, now in beta. You can now request fast mode the same way for every model on AI Gateway.
Vercel's AI Gateway now has one unified way to request fast mode, rather than provider-specific setup, in a change that's now in beta. Setting speed to fast in providerOptions.gateway upgrades a model to its fast serving path when one is routable, and falls back to standard speed when it isn't. Fast variants can also be addressed directly with a fast slug, such as anthropic/claude-opus-5-fast, which is handy in a gateway.models fallback list. Fast mode trades a higher per-token cost for lower latency or higher throughput, and only has an effect on models that offer a fast tier, marked with a lightning bolt in the models list.
In Claude Code, fast mode toggles with /fast for Anthropic Opus models after setup. Other coding agents can select a fast slug directly in their model configuration.
