Why is it 70% cheaper?
Flat monthly plans
You pay one predictable subscription, not per-token invoices that explode mid-month. Plans start at $6.99/mo and every plan includes every model in this graph.
Bulk-reserved compute
Luno reserves provider capacity in bulk at enterprise rates, then passes the discount on. That is how the same Claude Opus 4.8 costs 70% less here — same model, cheaper seat.
No markup on top of plans
The prices in this graph are effective rates inside your plan. No hidden per-request fees, no "premium model" surcharges, no surprise overage multipliers.
Model questions, answered
Yes. Every plan — from Starter to Dev — unlocks the full catalog: Claude Opus 4.8, GPT-5.5, Gemini 3.1 Pro and everything below them. Plans differ only in monthly usage volume, never in which models you can call.
Usually within 24–48 hours of the official release. The moment a provider ships a new model to their API, we add it to the gateway and it becomes available under the same key — no migration, no new endpoint.
Yes. Luno routes your requests to the same models with the same weights — we are a gateway, not a re-trained clone or a quantized mirror. Same outputs, same context windows, same capabilities, just a lower bill.
Change one string. Your key works for every model, so switching from GPT-5.5 to Claude Opus 4.8 is just editing the model field in your request body. Base URL stays https://api.luno.codes/v1 for everything.