What Pirol AI costs.

You buy budget and spend it on tokens. Let Pirol AI choose the model, or name one yourself — both are priced below.

Budget

from €0.62 / 1M output tokens

Buy credit, spend it on tokens
Every model, one endpoint
No subscription
From €20 — and it never expires

Enterprise

on request

Custom terms and volume pricing
SLAs
SSO
Integration support

No free tier and no subscription — you buy budget, and it is spent as you use it. Kyte is different: its Free plan comes with a monthly token allowance included — see kyte.pirol.ai/pricing.

What you can ask for

Model	Input	Output	Speed
Pirol AI picks one of the models below for you · one blended rate, whichever it chooses	€0.69	€2.55	—
GLM-5.3-Flash 320B MoE, 18B active: coding, agentic workflows, reasoning, tool use, long context, vision	€0.19	€0.62	349 tok/s
DeepSeek V4.1 Flash multimodal MoE: coding, reasoning, and tasks combining images and text	€0.40	€1.59	139 tok/s
GLM-5.3 750B MoE: coding, agentic workflows, reasoning, tool use, long context	€1.47	€5.45	455 tok/s

EUR per 1M tokens. All prices exclude VAT.

Looking for Kyte's plans? kyte.pirol.ai/pricing

> often asked

FAQ

Not for the Pirol API — you buy budget and spend it on tokens, with no subscription. Kyte is different: its Free plan includes a monthly token allowance.

You buy budget up front, from €20, and it is drawn down per token — input and output priced separately per model. Budget does not expire.

Custom terms and volume pricing, SLAs, SSO and integration support.

In the EU.