Skip to content
Graphene

Pricing

Your choice of frontier model — at 20% less

Run Claude Opus 4.8, GPT-5.5 and more through one API, billed per million tokens at a flat 20% under each provider's list price. No subscriptions, no per-seat fees, no minimums.

Your choice of frontier model

20% under list

Run the frontier models you already trust — billed per million tokens at 20% below each provider's list price. No tiers, no seats.

ModelContextInput / MOutput / M
Claude Opus 4.8
Anthropic
200k
$4.00
$5.00
$20.00
$25.00
GPT-5.5
OpenAI
400k
$3.20
$4.00
$12.80
$16.00
Claude Sonnet 4.6
Anthropic
200k
$2.40
$3.00
$12.00
$15.00
Gemini 3 Pro
Google
1M
$2.00
$2.50
$8.00
$10.00

How billing works

Pre-load credits and pay only for the input and output tokens you use. Spend and budgets are visible in your dashboard.

Every model is billed at 20% under its provider list price; your invoice is the sum of metered usage across the models you route to.

Enterprise & sovereign

Sovereign lanes (regional data residency), pinned models, dedicated capacity, and committed-use discounts are available under an enterprise agreement.

Learn about sovereign compute →

Questions, answered plainly

How am I billed?
Per token. You pay for the tokens you send and receive, at a flat 20% under each provider's list price, with the routed rate visible on every request in your dashboard. No subscriptions, no per-seat fees, no minimums.
Can I cap what a workspace spends?
Yes. Every workspace carries per-token metering and budgets, so you can set spend limits and watch usage against them in real time before a bill ever arrives.
Why is Graphene cheaper than going direct?
Graphene is an orchestration layer, not a resell margin. We route workloads across owned, partner and spot GPU capacity and pass the routing efficiency through as a flat discount. You keep OpenAI-compatible portability the whole time — switching away is the same one-line change as switching in.
Do you support data residency requirements?
Sovereign in-region routing is delivered as configuration under enterprise agreements, with Australian routing first on the roadmap. If you operate under APRA CPS 234 or comparable regimes, talk to us about a sovereign lane.

Get started

Create a workspace, issue your first API key, and route inference through Graphene with the same contracts you see in the docs.

Get startedRead the docs

Manage billing & credits · Browse models · Home