Cost calculator
What will your run cost?
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.
Your self-hosted cost
$2.875 per million tokens
$7.25/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 | 16.43× | 16,429 |
| mistral-small-4 mistral | $0.26 | 10.95× | 10,952 |
| gpt-5.6-luna openai | $0.45 | 6.39× | 6,389 |
| deepseek-v4-pro deepseek | $0.54 | 5.29× | 5,287 |
| mistral-large-3 mistral | $0.75 | 3.83× | 3,833 |
| gemini-3.5-flash-lite google | $0.85 | 3.38× | 3,382 |
| claude-haiku-4-5 anthropic | $2.00 | 1.44× | 1,438 |
| gemini-3.6-flash google | $3.00 | 0.96× | 958 |
| mistral-medium-3.5 mistral | $3.00 | 0.96× | 958 |
| grok-4.5 xai | $3.00 | 0.96× | 958 |
| gemini-3.5-flash google | $3.38 | 0.85× | 852 |
| claude-sonnet-5 anthropic | $4.00 | 0.72× | 719 |
| gemini-3.1-pro-preview google | $4.50 | 0.64× | 639 |
| gpt-5.6-terra openai | $4.50 | 0.64× | 639 |
| gpt-5.4 openai | $5.63 | 0.51× | 511 |
| claude-opus-4-8 anthropic | $10.00 | 0.29× | 288 |
| claude-opus-5 anthropic | $10.00 | 0.29× | 288 |
| gpt-5.6-sol openai | $11.25 | 0.26× | 256 |
| claude-fable-5 anthropic | $20.00 | 0.14× | 144 |
| 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).