What are routing preferences?

Most models on PayPerQ can be served by more than one upstream provider, and each provider trades off differently on speed, price, and reliability. Routing preferences let you steer that choice on a per-request basis by appending a :suffix to the model name — no separate endpoint, no config file, no code changes beyond the model string itself.

bash
Copy code
curl https://api.ppq.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-YOUR_API_KEY" \
  -d '{
    "model": "anthropic/claude-sonnet-4.6:nitro",
    "messages": [{"role": "user", "content": "What is the capital of France?"}]
  }'

That :nitro tells PayPerQ to prioritize the fastest provider for this one request. Drop the suffix and you get the default routing. That's the whole feature.

The suffixes

Suffix	Optimizes for	What it does
:nitro	Throughput	Routes to the highest tokens-per-second provider. Great for user-facing chat where latency is felt.
:floor	Cost	Routes to the lowest-priced provider. Great for batch jobs where cost is the constraint.
:exacto	Tool-calling quality	Prefers providers with the strongest tool-calling reliability signals instead of the cheapest. Great for agentic workflows.
:online	Fresh information	Enables web search for the request (adds relevant search results to the prompt).
:thinking	Reasoning	Selects a model's dedicated reasoning variant, when it has one (see the note below).
:extended	Context length	Selects an extended-context variant, where the model offers one.

Each is appended directly to the model slug:

text
Copy code
anthropic/claude-sonnet-4.6:floor      # cheapest provider
anthropic/claude-sonnet-4.6:exacto     # best tool-calling provider
openai/gpt-5.2:online                  # with web search

You can also use them on PayPerQ's short aliases — claude-sonnet-4:nitro works the same way as the full slug.

When to use which

• :nitro — interactive chat, live assistants, anything a human is waiting on.
• :floor — bulk processing, evals, background jobs, anything where you'd trade a little latency for a lower bill.
• :exacto — agents and tool-heavy workflows where a mis-formatted tool call costs you more than a few cents of routing.
• :online — questions about recent events, prices, docs, or anything past the model's training cutoff.

Good to know

Billing is unaffected — you're always charged the real cost. PayPerQ records the actual generation cost reported by the upstream after the request completes, so a suffix that routes you to a cheaper (:floor) or pricier-but-faster (:nitro) provider is reflected automatically in what you pay. There's nothing extra to configure.

:online costs extra — on paid requests. When web search runs, it bills a small per-request search fee on top of the model tokens. On free-tier models :online is a no-op: PayPerQ does not attach web search to free-tier requests, so there is no search and no fee.

Some models are pinned and ignore routing preferences. A handful of models are deliberately routed to a single provider — for example, low-latency "fast" variants and models pinned to a specific endpoint for cost or reliability reasons. On those, :nitro / :floor / :exacto have no effect because there's only one provider to choose from. If a routing preference doesn't seem to change anything, this is usually why.

:thinking and :extended are variant-specific. They only work on models that actually publish that variant. On a model that doesn't have one (Claude, for example), the request returns a no endpoints found error rather than silently ignoring the suffix. To turn on reasoning for models like Claude, use the reasoning request parameter — e.g. { "reasoning": { "effort": "medium" } } — instead of :thinking.

Unknown suffixes pass through untouched. If you append something PayPerQ doesn't recognize as a routing preference, the model string is forwarded to the upstream as-is — so a provider's own permaslug (for example one that already ends in :free) keeps working exactly as before.

Compatible with everything else

Routing preferences are just a model-string convention, so they work anywhere you already pass a model to PayPerQ — the chat completions API, your existing OpenAI-compatible SDK, or any tool that lets you set the model name. Nothing else about your request changes.

What is PayPerQ?

PayPerQ is the private way to use AI. You get the big-name models from Anthropic, OpenAI, Google and xAI, the best open-source models, and TEE models that run inside a hardware enclave where not even the model's host can see your data. Chat, image, video and audio, all in one place, with no account and no subscription. You top up a balance with crypto or a card, from as little as 10 cents, and pay only for what you use. Most users spend about $4 a month, against $20+ for a single ChatGPT subscription.

Privacy is built in rather than bolted on. Every chat request is served inside a secure enclave that PayPerQ's servers, staff and logs cannot read into, so PayPerQ never sees your prompts or responses. Choose a TEE model and the model's host cannot read them either.

Why Use PayPerQ?

• Private by default - chats are served inside a secure enclave PayPerQ cannot read, and every chat model shows its privacy tier (Anon, ZDR or TEE)
• No account needed - no sign-up and no email; a credit id is all it takes
• Pay with crypto or card - Bitcoin, Lightning, Monero, stablecoins and more, or a credit card
• Hundreds of AI models from every major provider in one place
• Pay per use - no subscriptions; top up from 10 cents and pay only for what you use, about 1 cent per query on average
• Your history stays yours - conversations are stored on your device by default

Getting Started

1. Visit ppq.ai
2. Top up your balance with crypto or a card (as little as 10 cents)
3. Pick a model, check its privacy tier, and start chatting

No account, no sign-up. Fund and go.