> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sailresearch.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Agent Cost Calculator

> Estimate what a long-running agent costs on Sail and compare against alternatives

You can estimate the inference costs of running your agent by picking a workload
profile below, or set your own per-turn token mix.

<div id="sail-cost-calculator" className="sail-calc">
  <p>
    The interactive calculator needs JavaScript, which your browser doesn't support. Here is a representative
    worked example instead: a deep-research agent that runs 50 turns and
    totals 4.0M cached input reads, 1.5M fresh input tokens, and 250K output
    tokens, priced on Sail's `zai-org/GLM-5.2-FP8` and on generic
    frontier-class tiers at list prices (Sonnet-class: \$3 input / \$15 output
    / \$0.30 cache reads per 1M tokens; Opus-class: \$5 / \$25 / \$0.50).
  </p>

  | Where it runs                             | Cost    | vs Sail `standard` |
  | ----------------------------------------- | ------- | ------------------ |
  | Sail `flex`                               | \$1.41  | 0.8x               |
  | Sail `standard`                           | \$1.86  | 1x                 |
  | Sail `priority`                           | \$2.51  | 1.3x               |
  | Sail `asap`                               | \$4.42  | 2.4x               |
  | Sonnet-class API, batch tier (24h window) | \$5.18  | 2.8x               |
  | Sonnet-class API                          | \$10.35 | 5.5x               |
  | Opus-class API                            | \$17.25 | 9.2x               |
</div>

## How the math works

```
cost per run = turns × (fresh × P_input + cached × P_cached + output × P_output) / 1,000,000
```

* **Fresh input**: tokens the model reads for the first time each turn (new
  tool results, search snippets, file contents).
* **Cached input**: tokens reread from prompt cache (the growing conversation
  history). Cache reads are billed at the `cached` rate.
* **Output**: tokens the model writes (reasoning and answers).

Sail rates come from the [Pricing](/pricing) page for the selected model and
[completion window](/completion-windows); daily and monthly figures multiply
by runs per day and a 30-day month.

## Assumptions and caveats

* **Frontier tiers are generic list prices.** "Sonnet-class" is \$3 input /
  \$0.30 cache reads / \$15 output per 1M tokens (batch tier = 50% off inside
  a 24-hour window); "Opus-class" is \$5 / \$0.50 / \$25.
* **The "same model, traditional provider" row is the model's ASAP price.** Sail
  sets its ASAP (always-on) rate to match what a traditional always-on provider
  charges for the same model, so it doubles as the "running this elsewhere"
  baseline, and the scheduled completion windows (`flex`, `standard`,
  `priority`) price below it. Models served only on ASAP show the row at \~1x;
  models with no ASAP tier use a fixed traditional-provider list price.
* **The model has to do the job.** For simplicity, the math assumes you're using a single frontier-class open
  model for your task. Often, we see the hybrid approach using both frontier closed models and open models, or a mix of open models.