NERIN
SHEET 03

Cost calculator

Same engine as the schedule

A headline rate per million output tokens is only half an answer — real work reads input too, and an agent adds its own turns on top. Set the shape of your workload below and the sheet recalculates, using the same cost model that prices every model in the schedule.

Files and diffs read in, a patch written out.

12:1
50M

Calculation note — Qwen3-Coder-480B-A35B, 12:1 input, direct

Output — 1M tokens$1.061
Input — 12M tokens at $0.265/M$3.184
Agent overhead — 25%$0.000
Per million output tokens$4.24

Monthly at 50M output

$212.24

Same volume, frontier list

$2550.00

12.0× cheaper than the same workload at published frontier closed-model list, saving $2337.76 a month at this volume.

Output and input rates are derived from the cost model on sheet 02. The frontier reference is published list for a closed model at $15.00/M output and $3.00/M input — a market observation, not our figure. Ratios are typical, not a measurement of your traffic.

ISSUE

Issue and revision

Bring a workload. We will show you what it costs here, and what it runs on, before you move it.

RevIssueHallsGPU
APod 01 energised, code hall only1384
BVision and knowledge halls in service31,152
CBiological design hall in service41,536