Pricing and the API wallet

One price, plan or API

The API costs the same as an Aderra plan: one price per model, whichever way you pay. There are no seats, no monthly fee and no minimum beyond a top-up.

What	Priced per
Chat	Million tokens: input, cached input and output each have their own price
Images	Image, per model
Video	Second of video, per model
Speech	1,000 billable characters of input (Chinese, Japanese and Korean characters count 3)
Web search (Messages)	Search

Chat prices

Per million tokens, in US dollars:

Model	Input	Cached input	Output
DeepSeek V4.1 Flash Unfiltered (`deepseek-v4.1-flash-uncensored`, the default)	$0.28	$0.07	$1.00
DeepSeek V4.1 Flash (`deepseek-v4.1-flash`)	$0.30	$0.006	$1.00
GLM-5.3 Flash (`glm-5.3-flash`)	$0.15	$0.03	$0.50
GLM-5.3 Flash Unfiltered (`glm-5.3-flash-uncensored`)	$0.25	$0.10	$1.00

These are the same on a plan. This table was updated on 2026-11-11; the authoritative figures are the live ones in GET /v1/usage/prices and on the Developers page.

Image, video and speech prices

Media prices are read live from GET /v1/usage/prices and shown in the Pricing section of the Developers page. They are not repeated here, because they change as Aderra’s own costs change.

Credits

Prices and balances are counted in credits: 1,000 credits = $1. The API reports exact amounts in micro-credits (1 credit = 1,000,000 micro-credits), as strings, so nothing is lost to rounding: billed_micro on an image, the x-aderra-billed-micro header on speech.

The API wallet

API keys pay from a prepaid API wallet, separate from any plan. API keys never spend plan credits, and a plan never spends the wallet.

Top up on the Developers page: Add funds, any amount from $5, then Pay by card or Pay with crypto. Credits arrive once the payment is confirmed.
API credits never expire.
The Developers page shows the available balance, what is held by calls in progress, spend by day, model and key, and every request with its tokens, cost and duration. It never shows what you sent or what the model answered.

If the page says “Buying API credits is not available on this account yet.”, contact us.

How a call is charged

Hold. Before the call runs, Aderra holds the most it could cost: for chat, your input plus the full max_tokens at the output price; for media, the price of the image, clip or text.
Settle. When it finishes, you pay what it actually used. The rest of the hold goes back to the wallet.
Failures are free. A call that fails, is refused by the model, or (for chat) produces no visible answer is not charged.

A nearly empty wallet can therefore refuse a chat request with a large max_tokens (402 api_wallet_empty) even though the answer would have cost less. Send a smaller max_tokens, or add funds. Each key’s monthly cap counts holds too.

Read prices from code

GET /v1/usage/prices returns every price, live, for any key. It is served on https://api.aderra.ai only. Amounts are credits as exact decimal strings.

curl https://api.aderra.ai/v1/usage/prices \
  -H "Authorization: Bearer $ADERRA_API_KEY"

import os
import requests

prices = requests.get(
    "https://api.aderra.ai/v1/usage/prices",
    headers={"Authorization": f"Bearer {os.environ['ADERRA_API_KEY']}"},
    timeout=30,
).json()

for m in prices["chat"]:
    print(m["model"], m["input_per_million"], m["output_per_million"])
for m in prices["image"]:
    print(m["model"], m["per_image"], m["sizes"])

const r = await fetch('https://api.aderra.ai/v1/usage/prices', {
  headers: { Authorization: `Bearer ${process.env.ADERRA_API_KEY}` },
});
const prices = await r.json();
for (const m of prices.chat) {
  console.log(m.model, m.input_per_million, m.output_per_million);
}

{
  "chat": [
    { "model": "glm-5.3-flash", "input_per_million": "…", "cached_input_per_million": "…",
      "output_per_million": "…", "max_output_tokens": 32768 }
  ],
  "image": [
    { "model": "…", "per_image": "…", "sizes": ["1024x1024"], "filtered": true, "default": true }
  ],
  "video": [
    { "model": "…", "per_second": "…", "min_seconds": 5, "max_seconds": 15,
      "requires_image": false, "explicit": false, "default": true }
  ],
  "speech": [
    { "model": "aderra-voice-1", "per_thousand_characters": "…", "max_characters": 4096,
      "voices": ["Reid", "Wesley", "…"] }
  ]
}

To turn credits into dollars, divide by 1,000.

A video model’s per_second is its price for a text or image clip. A reference clip costs more; every mode’s price is in the Video models table.