Vercel AI Gateway: Pricing, Models, Pros & Cons

Founded 2024|San Francisco, CA|201-500 people|Usage-based

What is Vercel AI Gateway?

Vercel AI Gateway is Vercel's managed gateway for calling hundreds of models across providers with one API key. It handles provider and model fallbacks, request logs, and spend budgets, and accepts AI SDK, OpenAI Chat Completions, OpenAI Responses, and Anthropic Messages requests. Apps don't need to run on Vercel to use it, although Vercel deployments can authenticate through OIDC instead of an API key.

Vercel charges no markup on tokens, including with your own provider keys, and bills against prepaid credits. The free tier covers a subset of models at lower rate limits, and BYOK requires buying credits first. Some controls cost extra once enabled, including team-wide zero data retention, custom reporting, and OpenTelemetry trace drains. Budgets also only apply to spend on Vercel's system credentials, so BYOK traffic sits outside those limits.

Observability covers latency, tokens, and spend per request, but Vercel AI Gateway doesn't score output quality. Teams that need evals on production traffic pair it with a separate tool or compare the Vercel AI Gateway alternatives that combine routing, tracing, and evals in one platform.

Key features

Core capabilities this platform advertises.

One API key for hundreds of models
Provider and model fallbacks
Request logs and spend dashboards
Budgets by team, project, API key, or member

Strengths and tradeoffs

What this tool does well, and the limitations to keep in mind.

Pros

No markup on tokens, including with BYOK
Callable from any infrastructure, not only Vercel deployments
Ordered provider and model fallbacks

Cons

No output evaluation or quality scoring
Free tier limited to a subset of models with lower rate limits
Budgets don't cover BYOK spend, and failed BYOK requests fall back to billed system credentials
Team-wide ZDR, custom reporting, and trace drains are paid add-ons

Plans & pricing

What's included in each plan, and how the tiers compare.

Free Tier

Monthly free credit
Subset of models
Lower per-model rate limits

Pay-As-You-Go

Provider list price

Per token

Zero markup, including BYOK
Higher rate limits
Payment processing fees on credit purchases
Team-wide ZDR $0.10 per 1k requests (Pro and Enterprise)
Trace drains $0.05 per 1k traces (Pro and Enterprise)
Custom reporting $0.075 per 1k writes, $5 per 1k queries

Enterprise

Custom

Contact sales for a quote

Invoiced billing with no processing fees
Volume discounts on token spend

View official pricing page

Common use cases

Developers who want managed multi-provider model access and spend controls, especially teams already on Vercel or the AI SDK

Multi-provider model access for AI SDK apps
Provider failover for production agents
Capping AI spend by project or API key
Routing coding agent spend through one account

Using Vercel AI Gateway with Respan

Respan traces Vercel AI SDK apps through OpenTelemetry with @respan/exporter-vercel, capturing full input and output, token usage, and cost for every call. Respan can also serve as the gateway for AI SDK apps: point the OpenAI provider at Respan's endpoint to route across 1,000+ models, and each request is routed, traced, and scored on the same record.

Trace AI SDK calls with input, output, tokens, and cost
Route across 1,000+ models through one OpenAI-compatible endpoint
Score output on live production traffic with automated evals
Tag traces by customer, thread, or workflow

Route, trace, and score every model call

10KFree traces/mo

1,000+Models

5 minSetup

Respan routes across 1,000+ models, traces every request, and scores output on live traffic, so cost, latency, and quality sit on one record. Start free.

Try Respan free

Best Vercel AI Gateway alternatives & competitors

Top companies in LLM Gateways you can use instead of Vercel AI Gateway.

Respan

One API for 1,000+ models