Pricing / Early access
A pilot scoped
to your workload.
Commercial pricing is not published yet. We’ll agree workload limits, support and commercial terms with you before a pilot begins.
Understand the cost
before you commit.
Share your expected usage, the models you’re considering and the support your team needs. We’ll use those requirements to discuss a suitable pilot.
Enquiries are free. Requesting access does not create an account or incur a charge.
From the prototype catalog
Model reference rates
USD per one million tokens
| Model | Input | Output |
|---|---|---|
| GPT-5 | $5.00 | $15.00 |
| GPT-5 Mini | $0.40 | $1.60 |
| GPT-5 Nano | $0.05 | $0.40 |
| Claude Opus 4.8 | $6.00 | $30.00 |
| Claude Sonnet 4.6 | $3.00 | $15.00 |
| Claude Haiku 4.5 | $0.80 | $4.00 |
| Gemini 2.5 Pro | $2.50 | $10.00 |
Reference values only. Not a commercial quote or a live price feed. Provider availability is configuration-dependent.
View the full catalogWhat we’ll discuss
Scope, usage and support.
Early access
Get familiar with the API and the product direction before a production pilot.
- OpenAI-compatible interface
- Prototype model catalog
- Developer documentation
Usage, made visible
The goal: understand what each request costs, with controls that keep spending intentional.
- Versioned model pricing
- Auditable usage and charges
- Project and application budgets
Enterprise control
A path to your own provider contracts and approved infrastructure, as the foundations mature.
- Bring your own provider account
- Regional policy controls
- Private inference endpoints
Pricing questions
Before your pilot.
Are these final commercial prices?
No. The rates shown here come from the prototype model catalog. They are reference values, not a quote or a live provider price feed. Final commercial terms will be validated before a production pilot.
Can I use RouterCore in production today?
RouterCore is a development preview. Production access follows the engineering review gates for provider reliability, policy, metering, billing and security.
How will usage be charged?
The planned model is usage-based, with input and output tokens priced separately. Correct metering, an auditable ledger and protection against duplicate charges come before broader commercial rollout.
Will I be able to set a budget?
Project and application budgets are part of the product plan. Policy and budget enforcement must pass their review gate before they are presented as production guarantees.