Vercel AI Gateway vs Standard Compute
Vercel AI Gateway sells tokens at list price with no markup and leaves the model choice to a classifier you build, while Standard Compute does the cost optimization for you: its router picks a cost-efficient model per request inside a flat monthly plan, which suits teams that want lower agent bills without writing routing code.
Side by side
| Vercel AI Gateway | Standard Compute | |
|---|---|---|
| Pricing | Provider price with no markup or platform fee; payment fees; prepaid credits expire after one year | Flat monthly plan: $19, $39, $89 or $249 with $20 to $269 of compute included; Business from $499 |
| Who picks the model | You pick, or order your own candidates; no built-in classifier | Router picks per request (harder steps to stronger models, routine steps to efficient ones), or pin any model ID |
| Budget control | Budgets per team, project, key and member; Vercel calls them a soft cap | Plan budget per key; new requests get HTTP 402 when it is used, no overage |
| Claude Code | Native | Native Messages API |
| Codex | Native | Native Responses API, any model in the catalog |
| Europe / US rules | Inference region per request | Opt-in Europe or US rules on every plan, per key, team or company |
| Teams | Vercel teams and projects | Company and agency workspaces, a plan per key |
| Best for | Teams on Vercel who want list-price tokens | Coding agents and agent fleets that need one fixed monthly cost |
Vercel AI Gateway facts from its own pricing and docs pages. Standard Compute is our product.
Where they differ
- Vercel sorts providers for the model you name and runs routers over your own candidate list; its cost-routing guide says the classifier is the part you own. Standard Compute's router classifies each request and picks the model.
- Vercel charges provider prices with no markup or platform fee, plus payment fees, from prepaid credits that expire after a year. Standard Compute is a flat plan with the compute included.
- Vercel sets the inference region per request. Standard Compute's Europe or US rules are set once for a key, team or company and enforce zero data retention on every request while they apply.
When Vercel AI Gateway is the better fit
Vercel AI Gateway is the better pick if you already build on Vercel, want list-price tokens with fine-grained budgets per project and member, or want its dedicated Claude Code and Codex endpoints.
Switch from Vercel AI Gateway in 2 minutes
Standard Compute speaks the same APIs, so switching is a base URL, a key and a model name. Your prompts, tools and streaming stay as they are.
| Setting | Vercel AI Gateway | Standard Compute |
|---|---|---|
| OpenAI-compatible base URL | https://ai-gateway.vercel.sh/v1 | https://api.stdcmpt.com/v1 |
| Claude Code base URL | https://ai-gateway.vercel.sh/claude-code | https://api.stdcmpt.com |
| API key | Your Vercel AI Gateway key | Your Standard Compute key |
| Model | The model ID you use today | standardcompute, or any exact model ID |
# OpenAI-compatible clients (OpenCode, Cline, Kilo Code, SDKs)
base_url = "https://api.stdcmpt.com/v1"
api_key = "<your-standard-compute-key>"
model = "standardcompute" # or pin one, e.g. anthropic/claude-sonnet-5.5# Claude Code (Anthropic Messages API, no /v1)
export ANTHROPIC_BASE_URL="https://api.stdcmpt.com"
export ANTHROPIC_AUTH_TOKEN="<your-standard-compute-key>"# Codex CLI, ~/.codex/config.toml
model = "standardcompute"
model_provider = "standardcompute"
[model_providers.standardcompute]
name = "Standard Compute"
base_url = "https://api.stdcmpt.com/v1"
env_key = "STANDARD_COMPUTE_API_KEY"
wire_api = "responses"Run your agents on a fixed monthly budget
Every plan includes the compute and every model, and the router spends the budget on the most cost-efficient model for each step. Pay-as-you-go top-ups are optional, never automatic.
FAQ
Is Vercel AI Gateway cheaper than Standard Compute?
Per token, Vercel charges exactly the provider's price. Standard Compute includes 5% to 14% more compute than its plan price and saves through routing routine steps to efficient models, so the cheaper option depends on how much of your work an efficient model can do and whether you use your plan's budget.
Does Vercel AI Gateway choose models automatically?
Not by itself. Vercel can sort providers for one model by cost or speed and order your own list of candidates, and its guide on cost-aware routing says the classifier is the part you own. Standard Compute's router picks the model per request when you send model standardcompute.
Are Vercel AI Gateway budgets hard limits?
Vercel's docs call a budget a soft cap: the request that crosses the limit still completes. Standard Compute behaves similarly at the edge (running requests finish) and answers every new request with HTTP 402 once the plan budget is used.
Do Vercel AI Gateway credits expire?
Yes. Vercel's FAQ says purchased credits expire one year after purchase. A Standard Compute plan budget renews each billing period instead.
How do I run Claude Code on each?
On Vercel, set ANTHROPIC_BASE_URL to https://ai-gateway.vercel.sh/claude-code with your gateway key in ANTHROPIC_AUTH_TOKEN. On Standard Compute, set ANTHROPIC_BASE_URL to https://api.stdcmpt.com, or run standardcompute launch claude.
Sources: Vercel AI Gateway pricing · Vercel AI Gateway budgets · Vercel: cost-aware model routing · Vercel AI Gateway with Claude Code