Inference Cost Calculator
Calculate the cost of serving LLM inference at scale with self-hosted GPUs vs. API providers.
Self-Hosted Results
GPUs Needed
2
Min 2 for VRAM
Cost per 1K Input Tokens
$0.0240
Daily Cost
$168.00
Monthly Cost
$5.0K
vs. API Providers
| Provider | Input $/1K | Output $/1K | Daily Cost | Monthly Cost | vs. Self-hosted |
|---|---|---|---|---|---|
| Self-hosted (H100 SXM) | $0.0240 | $0.0240 | $168.00 | $5.0K | -- |
| OpenAI GPT-4o | $0.0025 | $0.0100 | $32.50 | $975.00 | -81% |
| Anthropic Claude Sonnet | $0.0030 | $0.0150 | $45.00 | $1.4K | -73% |
| Google Gemini Pro | $0.0013 | $0.0050 | $16.25 | $487.50 | -90% |