Rent the GPU, or pay per token?

The two sides of this site, bridged: today’s GPU rental medians against the labs’ token list prices. We supply both price sides live; the throughput is yours to supply, because it varies enormously with model, quantization, and batching — and we don’t print performance numbers we can’t stand behind. The arithmetic is shown under every result.

GPU (today’s median)Your throughput (tok/s per GPU)Example figure — replace with YOUR measured, batched throughput. It varies 10×+ with model, quantization, context, and batching.Utilization: 70%Fraction of rented hours doing useful work.

Your self-hosted cost

$2.933 per million tokens

$7.39/hr × 1,000,000 ÷ (1,000 tok/s × 3600s × 70%)

API model	Blended $/MTok	Self-host vs API	Break-even tok/s
deepseek-v4-flash deepseek	$0.18	16.76×	16,757
mistral-small-4 mistral	$0.26	11.17×	11,172
gpt-5.6-luna openai	$0.45	6.52×	6,517
deepseek-v4-pro deepseek	$0.54	5.39×	5,393
mistral-large-3 mistral	$0.75	3.91×	3,910
gemini-3.5-flash-lite google	$0.85	3.45×	3,450
claude-haiku-4-5 anthropic	$2.00	1.47×	1,466
gemini-3.6-flash google	$3.00	0.98×	978
mistral-medium-3.5 mistral	$3.00	0.98×	978
grok-4.5 xai	$3.00	0.98×	978
gemini-3.5-flash google	$3.38	0.87×	869
claude-sonnet-5 anthropic	$4.00	0.73×	733
gemini-3.1-pro-preview google	$4.50	0.65×	652
gpt-5.6-terra openai	$4.50	0.65×	652
gpt-5.4 openai	$5.63	0.52×	521
claude-opus-4-8 anthropic	$10.00	0.29×	293
claude-opus-5 anthropic	$10.00	0.29×	293
gpt-5.6-sol openai	$11.25	0.26×	261
claude-fable-5 anthropic	$20.00	0.15×	147
gpt-5.5-pro openai	$67.50	0.04×	43

Below 1.00× self-hosting is cheaper per token than that API at your inputs; the break-even column is the throughput where they match. This compares raw token economics only — it ignores your engineering time, the API’s quality/latency, and that a rented GPU bills whether or not you keep it busy (that’s the utilization slider).