AI API cost, measured

Not list prices — billing data. Every analysis below is computed from real calls served by the gateway and can be recomputed from the public price feed.

Cheapest text run
FREE
GLM-4.7-Flash · 1M in + 200K out
Cheapest 5s video
$0.154
PixVerse C1 Video API · per clip
Cheapest image
FREE
CogView3-Flash · per image

The series

75.9%
Which model took 41% of the bill
upper bound if every call used the cheapest model
0.005%
Audit: 0% markup verified
deviation from provider list price across 58 models
$0.0048
What real calls cost
median paid call, 13,807 calls
$0.308
What media generation costs
median 5s video clip, 121 calls
324ms–68s
Latency by model, measured
30 days of gateway traffic, 63 models
Prices and cheapest tiles reflect the live feed at page generation time. The written analyses cover the 30-day window ending 2026-10-22.