fim@gate:~$ fim pricing ls
Model Pricing
Live rates pulled straight from the gateway — what you see here is what billing uses. Token prices in USD per 1M tokens.
fim pricing ls --live
0 / 0 modelsvendor:
billing:
group:
modelvendorinput $/1Moutput $/1Mcachedgroup
no models match — try a different pattern
token prices are per 1M tokens; per-request models are billed per call.
"best price" shows each model at its cheapest available channel group; pick a group to lock all rows to it.
cached = price for input tokens served from prompt cache; cache writes (Anthropic models) follow the official rule — 5m TTL = input ×1.25, 1h TTL = input ×2.
rates update automatically when upstream pricing changes.