Question 1

What am I paying for?

Accepted Answer

Your plan buys a number of priority concurrent requests, plus access to overflow capacity. Included inference has no monthly token cap or per-token overage charges. External providers and optional tools are priced separately.

Question 2

What does private mean?

Accepted Answer

Inference prompts and responses are discarded after processing and aren’t used for training. We keep account, usage, and billing metadata. Subscriptions share infrastructure; Enterprise offers dedicated deployments.

Question 3

Which models are included?

Accepted Answer

Coding plans include the open-weight models listed above, served by LunaRoute on our own GPU fleet. Connected external providers are billed separately.

Question 4

Is there a monthly token limit?

Accepted Answer

Included inference has no monthly token cap. Your plan sets concurrent capacity and queue limits; scheduling, timeouts, and service terms still apply.

Question 5

What happens above my included concurrency?

Accepted Answer

Additional requests may run as overflow using available capacity, without per-token overage charges for included inference. They may take longer to start during busy periods. Overflow is not unlimited: plan capacity, queue limits, and timeouts apply, and requests may be rejected when those limits are reached.

Question 6

Does priority mean a guaranteed speed?

Accepted Answer

No. Priority means preferential scheduling, not a guaranteed response time or tokens-per-second rate.