Tool · Run a model

Self-host vs API — at what volume does renting a GPU win?

Open-weight models can run on rented GPUs (~$1–$4/hr for an A100 / H100) or via per-token APIs. Plug in your numbers; we'll show the breakeven curve.

Monthly input tokens

30,000,000

@ 1,000,000/day

Monthly output tokens

15,000,000

@ 500,000/day

Self-host GPU

Nvidia GeForce RTX 5090

needs 25 GB VRAM

Self-host monthly

$327

24/7 on cheapest provider

API pricing — cheapest first

Monthly bill at 30,000,000 input + 15,000,000 output tokens.

Provider	$/M in	$/M out	Monthly	vs self-host
Alibaba DashScope	$0.04	$0.15	$3.45	−$323.79	Open →
OpenRouter	$0.1	$0.15	$5.25	−$321.99	Open →
DeepInfra	$0.1	$0.15	$5.25	−$321.99	Open →
Together AI	$0.17	$0.25	$8.85	−$318.39	Open →

Cheapest fitting GPU for Qwen: Qwen3.5-9B at FP16. Live at Nosana at $0.455/hr → $327/month at 24/7 utilisation.