← Sodura

Pricing

Simple, transparent pricing

Pay only for what you use. No monthly seats, no hidden fees. Billed per million tokens.

🎉 Free beta through July 15, 2026 — use it free, pricing kicks in after.
📥

Input tokens

Messages you send to the model

500

per 1 million input tokens

100K tokens₮50
1M tokens₮500
10M tokens₮5,000
📤

Output tokens

Responses generated by the model

5,000

per 1 million output tokens

100K tokens₮500
1M tokens₮5,000
10M tokens₮50,000

All models, same pricing

Every model tier uses the rates above. Choose based on speed and quality, not cost tiers.

ModelSpecsBest forInput / 1MOutput / 1M
sodura-smallGemma 12B · 8-bitHigh-volume, low latency₮500₮5,000
sodura-mediumGemma 26B · 4-bitGeneral purpose (default)₮500₮5,000
sodura-largeGemma 31B · 4-bitHighest capability₮500₮5,000

Common questions

What is a token?

A token is roughly 4 characters of text (~¾ of a word). A typical support ticket reply is around 200–400 tokens.

When does billing start?

Billing is deferred during the free beta (through July 15, 2026). Usage is tracked but not charged. Pricing listed here applies after the beta ends.

Is there a free tier after beta?

A free trial credit will be available after the beta. Details will be announced before the beta ends.

Do input and output tokens cost the same?

No. Input tokens (your messages) cost ₮500/M. Output tokens (model responses) cost ₮5,000/M — 10× more, reflecting generation compute.

What currency is this?

Prices are in Mongolian Tögrög (₮). If you need USD or other currency billing, contact us for enterprise options.

Is there an enterprise plan?

Yes. Contact eba@sodura.ai for volume discounts, SLA, private deployments, and custom model options.

Start free today

No credit card needed during beta.