Compare
Cent AI vs OpenRouter
Both are one API in front of many models. This page puts the differences side by side; every OpenRouter cell links to OpenRouter's own page and the date it was read.
In three lines
- 1Prices: OpenRouter bills the hosting provider's list price with no markup and adds a fee when you buy credits; Cent AI sets its own per-token rate per model and shows it next to the official price.[1][2]
- 2Payment: Cent AI takes QRIS in rupiah and USDT/USDC; OpenRouter lists cards, AliPay and USDC.[1]
- 3Catalog: OpenRouter lists far more models and providers; Cent AI keeps a short list with four published rates per model.[2]
Side by side
- Token prices
- Cent AI: Own per-token rate per model, shown next to the official price, with the price version and its date. Each successful request costs at least $0.00002.
- OpenRouter: The hosting provider's list price, with no markup on inference.[1][2]
- Fee on buying credit
- Cent AI: None added: Cent AI pays the payment fees, and QRIS uses the live exchange rate with no spread.
- OpenRouter: 5.5% per credit purchase, $0.80 minimum, on the Standard plan; 8% on Business; 5% for crypto payments.[1][2]
- Payment methods
- Cent AI: QRIS in rupiah, USDT and USDC on Solana.
- OpenRouter: Cards, AliPay and USDC; Enterprise by invoice. QRIS is not among the listed methods.[1][2]
- Protocols
- Cent AI: OpenAI Chat Completions and Responses, Anthropic Messages.
- OpenRouter: OpenAI Chat Completions and Completions, a stateless Responses API, Anthropic Messages.[1][10][9]
- Trial balance
- Cent AI: $1 promo balance for the first 100 accounts, then $0.10; once per account, all models, 30 days.
- OpenRouter: A small free allowance for new users, plus free models limited to 50 requests a day (1,000 after buying 10 or more credits).[1][3]
- Key and project controls
- Cent AI: Daily and monthly limits per key, a budget per project, and IP and model allowlists per key.
- OpenRouter: Optional spend cap per key, workspace budgets and guardrails.[3][4]
- Cost per request
- Cent AI: usage.cost in every response, and in the last chunk of a stream.
- OpenRouter: usage.cost in every response, and on the Activity page.[5]
One workload, priced
100 million tokens a month, 99% input, 1% output, no cache. Cent AI from the active price version; OpenRouter from its published price for the model.
- claude-sonnet-5.571% off
Price version v6 · Oct 31, 2026
Official priceOfficial price $208.00
Cent AI$60.32
OpenRouter: list price × workload, plus the 5.5% fee on buying credit (Standard plan, card; crypto 5%), for one purchase of $14.55 or more, below which the $0.80 minimum applies. For this model, regional endpoints cost 10% more.[1][2][11]
Open-weight models are left out: OpenRouter routes them across hosts with different prices, so one number would mislead.
When OpenRouter fits better
You need a model Cent AI does not list
OpenRouter's pricing page counts 500+ models from 80+ providers.[2]
You already hold provider keys
Bring your own keys: on pay-as-you-go the first $25,000 of list-price inference a month carries no BYOK fee, then 5%.[1][2]
You want free models to try
OpenRouter lists 25+ free models, with low daily limits.[2][1]
Your company buys by invoice
OpenRouter's Enterprise plan offers invoicing and purchase orders.[2]
Moving from OpenRouter
Change the base URL and the key. Model IDs drop the maker prefix: z-ai/glm-5.3 becomes glm-5.3.[7][8]
import os
from openai import OpenAI
# Before: OpenAI(base_url="https://openrouter.ai/api/v1", api_key=os.environ["OPENROUTER_API_KEY"])
client = OpenAI(base_url="https://api.cent-ai.dev/v1", api_key=os.environ["CENT_AI_KEY"])
resp = client.chat.completions.create(
model="glm-5.3", # before: "z-ai/glm-5.3"
messages=[{"role": "user", "content": "Hello"}],
)
print(resp.usage)
Request parameters reach the model as you send them.
Sources
- [1]https://openrouter.ai/docs/faq (opens in a new tab) · read Oct 29, 2026
- [2]https://openrouter.ai/pricing (opens in a new tab) · read Oct 29, 2026
- [3]https://openrouter.ai/docs/api_reference/limits (opens in a new tab) · read Oct 26, 2026
- [4]https://openrouter.ai/docs/guides/best-practices/spend-controls (opens in a new tab) · read Oct 26, 2026
- [5]https://openrouter.ai/docs/cookbook/administration/usage-accounting (opens in a new tab) · read Oct 26, 2026
- [6]https://status.openrouter.ai/ (opens in a new tab) · read Oct 26, 2026
- [7]https://openrouter.ai/docs/quickstart (opens in a new tab) · read Oct 26, 2026
- [8]https://openrouter.ai/docs/cookbook/coding-agents/claude-code-integration (opens in a new tab) · read Oct 26, 2026
- [9]https://openrouter.ai/docs/api/api-reference/anthropic-messages/create-a-message (opens in a new tab) · read Oct 26, 2026
- [10]https://openrouter.ai/docs/api_reference/responses/overview (opens in a new tab) · read Oct 26, 2026
- [11]https://openrouter.ai/api/v1/models/anthropic/claude-sonnet-5.5/endpoints (opens in a new tab) · read Oct 26, 2026
OpenRouter facts are re-read every month; a sentence whose source changed is taken down.
Try Cent AI on your own workload
Sign in with GitHub or Google and call any model with the promo balance.