Standard Compute

Which API should you use with Amp?

To put Amp's main agent on one monthly price, add Standard Compute as a Custom URL connection in Amp's Model Routing (base URL https://api.stdcmpt.com/v1) and pin your modes to models it serves, from $19 a month with a $20 compute budget. Without a connection of your own, Amp bills model usage from credits at providers' API prices, so a long high or ultra thread costs what the API would. Custom URL connections work on every Amp tier, including the free Hobby plan.

Guide updated 2026-10-14. Check each provider's current terms below.
Light or occasional use
For a few threads a week, Amp's own credits at API price, or a ChatGPT subscription linked in Model Routing, will likely cost the least. Amp's low mode already runs its main agent on an efficient model.
One fixed monthly bill
When medium and high threads on Claude Opus or GPT-6 Astra become a visible line on your Amp credits, map those models to Standard Compute and pin them in Tune Modes. At the monthly budget Standard Compute answers HTTP 402 instead of billing more, and anything you leave unmapped stays on Amp.

API and subscription options for Amp

OptionPriceWhat you get
Amp creditsPay-as-you-goAmp's own model access. For individuals, paid credits are charged at providers' API prices with no markup, need no paid tier and expire 12 months after purchase. Hobby is free; the $20 Individual tier adds orb time. Source: ampcode.com
Standard Compute (us)From $19/moStandard Compute starts at $19/month with a $20 monthly compute budget. Requests stop when that budget is exhausted until renewal or a plan change; there are no automatic overage charges. Requests are not slowed as the budget is used. Pin any model by its exact ID, or let smart routing pick per request. It is a fixed bill with a finite allowance.
OpenRouterPay-as-you-goChoose models and providers through one API. Usage-based pricing can suit light or bursty work; estimate long agent sessions and review current fees and limits. Source: openrouter.ai
Anthropic APIPay-as-you-goDirect API access for supported Claude models. Token use, model choice, service limits and spending controls determine the cost and available throughput. Source: claude.com

Prices and allowances can change. Amp's tiers and bring-your-own-key terms were read from ampcode.com/pricing and Amp's Pricing and Model Routing docs on October 14, 2026. Amp pricing · Amp: model routing. Spot an error? Tell us.

Add Standard Compute in Amp's Model Routing

Settings → Model Routing → Add → Custom URL
Format         chat-completions
Base URL       https://api.stdcmpt.com/v1
API key        sc_live_…            (your Standard Compute key)
Model mapping  openai/gpt-5.6-sol
               moonshotai/kimi-k3
               # Claude: pick it in the Models picker, then add
               # -> anthropic/claude-opus-5.5 (Standard Compute uses dots)

Press Check Access, then pin the roles you want in Settings → Mode Dial → Tune Modes. Web search, image generation, voice and orb minutes stay on Amp credits.

Amp questions

Does Amp's free plan work with my own OpenAI-compatible endpoint?
Yes. Amp's docs list AI routers, gateways and any OpenAI- or Anthropic-compatible endpoint as Custom URL connections for all users on every tier, and the pricing page lists bring-your-own-key on the free Hobby plan with no Amp token fees. Add the connection under Settings → Model Routing, map the models it should serve, and pin those models in Tune Modes. Roles you do not pin keep Amp's defaults and bill Amp credits.

Try one task before choosing a plan

  1. Check the API format and model configuration supported by your installed Amp version.
  2. Run a small task on a clean branch and review the result, elapsed time and compute used.
  3. Estimate your actual monthly workload before selecting a larger allowance.
Get your API keyAmp setup guide →Compare compute budgets →

Record your own results with the workload CSV and report template. These are blank templates, not published benchmark results.

Disclosure: this guide is maintained by Standard Compute, one of the options above. Choose based on task quality, supported features and total cost for your workload. Light usage may cost less on a usage-based API; sustained workloads need enough allowance on any subscription.

All flat-rate plans compared →The cheapest-tokens math →Plan compute for your team →