Pricing / AI usage

AI usage and rates.

Each plan includes a monthly AI usage budget. The token rates below explain how much of that budget a request uses.

← Compare plans

Model token rates

USD per million tokens, at standard context. Model availability is shown in the app; a rate entry does not enable a model.

Rates used to express the included AI usage budget
ModelInputCached inputOutput, including reasoning
gpt-5.6-luna$0.20$0.02$1.20
gpt-5.6-terra$2$0.20$12
gpt-5.6-sol$4$0.40$20
gpt-6-luna$0.10$0.01$0.50
gpt-6.1-terra$1$0.05$5
gpt-6.1-sol$2$0.10$10
gpt-6-astra$10$1$50

Rate version: sf-service-usd-2026-11-13. Reviewed against OpenAI’s model pricing. The monthly subscription covers SiliconForge’s application and managed service; this table is not an additional bill.

One example.

At standard context, 100,000 uncached input tokens and 10,000 output tokens use $0.015 of the AI budget with GPT-6 Luna, or $0.30 with GPT-6.1 Sol. A task can require several requests.

A monthly allowance

The budget refreshes each month, including on annual billing. It is shared across your team, does not roll over and cannot be withdrawn as cash.

No automatic overage

We reserve enough budget before generation starts, then settle the returned usage. If you cannot cover the reservation, the request does not start. An unconfirmed request may remain reserved while its usage is reconciled; contact support rather than repeatedly resending it.

Local checks are separate

Lint, simulation, synthesis and packaging do not consume managed AI budget. They still require valid application access. External AI accounts have their own provider billing.

Long-context and cache-write rates

Above 272,000 input tokens, input and cached-input rates double and output rates multiply by 1.5. When reported by the provider, 30-minute cache writes use 1.25× the normal input rate and 12-hour writes use 2.5×. Unsupported models or processing tiers are not silently substituted.

Existing subscription balances

Existing prices and balances are unchanged. The plan and account views express usage at the model token rates above. Older unit-based plans retain their original allowance. Amounts in the account view are rounded to cents for display.