FAQ
Common questions about accounts, models and billing.
Can I switch models without changing my code?
Yes. Model access is unified, so switching is a single model parameter — no SDK or integration changes.
Do you support function calling, vision and JSON mode?
These follow the capabilities of the upstream model you select. Pass the corresponding fields exactly as you would with the OpenAI API.
How do I know which models I can use?
Call GET /v1/models, or open Model Square in the console to see model names, groups and prices.
Where do I see what a request cost?
Open Usage Logs. Each entry lists the model, input/output tokens, the quota charged and the latency.
My API key leaked — what now?
Delete the key in API Keys and create a new one. Revocation takes effect immediately, and the new key works right away.
Requests fail with 402 but my balance looks fine
Check whether the model belongs to a group with a separate quota, and whether the request itself is very large (long context is billed as input tokens). Details are in Usage Logs.
Do you support self-hosted or custom models?
Availability depends on the deployment configuration. If a model is missing from Model Square, ask support whether it can be enabled.