Cloudflare AI Gateway vs Standard Compute
Cloudflare AI Gateway is a free control layer with optional Unified Billing at a 5% credit fee, while Standard Compute routes each request to a cost-efficient model inside a flat monthly plan, which suits agent fleets that want routing and a fixed cost without building flows or hitting a 200-requests-a-minute ceiling.
Side by side
| Cloudflare AI Gateway | Standard Compute | |
|---|---|---|
| Pricing | Gateway free; Unified Billing adds 5% to credit purchases; or bring your own keys | Flat monthly plan: $19, $39, $89 or $249 with $20 to $269 of compute included; Business from $499 |
| Who picks the model | Dynamic Routing flows you configure | Router picks per request (harder steps to stronger models, routine steps to efficient ones), or pin any model ID |
| Budget control | Spend limits per gateway, scoped by model, provider or metadata | Plan budget per key; new requests get HTTP 402 when it is used, no overage |
| Claude Code | Documented through the gateway's Anthropic path | Native Messages API |
| Codex | OpenAI models only | Native Responses API, any model in the catalog |
| Europe / US rules | Not documented for AI Gateway | Opt-in Europe or US rules on every plan, per key, team or company |
| Teams | Cloudflare account members | Company and agency workspaces, a plan per key |
| Best for | Teams already on Cloudflare who bring their own keys | Coding agents and agent fleets that need one fixed monthly cost |
Cloudflare AI Gateway facts from its own pricing and docs pages. Standard Compute is our product.
Where they differ
- Cloudflare's Dynamic Routing runs the conditions, splits and fallbacks you configure. Standard Compute picks the model per request.
- Unified Billing is limited to 200 requests per 60 seconds per gateway. Standard Compute's ceiling is per key: 300 requests a minute on Starter, rising to 2,000 on Pro.
- Cloudflare's Codex setup works only with OpenAI models that support the Responses API. Standard Compute serves Codex with any model in its catalog.
When Cloudflare AI Gateway is the better fit
Cloudflare AI Gateway is the better pick if you bring your own provider keys and want its free caching, logging and edge security around them.
Switch from Cloudflare 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 | Cloudflare AI Gateway | Standard Compute |
|---|---|---|
| OpenAI-compatible base URL | https://gateway.ai.cloudflare.com/v1/<ACCOUNT_ID>/<GATEWAY_ID>/openai | https://api.stdcmpt.com/v1 |
| Claude Code base URL | https://gateway.ai.cloudflare.com/v1/<ACCOUNT_ID>/<GATEWAY_ID>/anthropic | https://api.stdcmpt.com |
| API key | Your Cloudflare 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
How much does Cloudflare AI Gateway cost?
The core gateway is free. Unified Billing, where Cloudflare pays the providers for you, adds a 5% fee to credit purchases and passes provider rates through. You can also bring your own provider keys.
What is Cloudflare AI Gateway's rate limit?
Unified Billing is limited to 200 requests per 60 seconds per gateway and returns 429 above that; requests on your own keys are not counted. Standard Compute's ceiling is per key, from 300 requests a minute on Starter to 2,000 on Pro.
Can Codex use Claude through Cloudflare AI Gateway?
No. Cloudflare's Codex guide says only OpenAI models that support the Responses API work; Anthropic and Google models do not. Standard Compute serves Codex with any model in its catalog.
Does Cloudflare AI Gateway route to cheaper models automatically?
Not on its own. Dynamic Routing follows the rules you write, and spend limits can fall back to a cheaper model you name. Standard Compute's router picks a cost-efficient model per request.
Can I use both?
Yes. Standard Compute is OpenAI-compatible, so any gateway or client that accepts a custom base URL can send requests to https://api.stdcmpt.com/v1.
Sources: Cloudflare Unified Billing · Cloudflare AI Gateway limits · Cloudflare AI Gateway with Codex