LLM Cost & Latency Calculator
Pick a model and your usage. Cloud API shows what you will spend per month on tokens; on-prem recommends a GPU and shows the latency your users will see.
1. How will you run the model?
2. Choose the LLM
$0.15 input / $0.6 output per 1M tokens (list price, Sep 2026) · includes ~50% extra hidden reasoning tokens
3. Pyxon Eye
4. Users and sessions
GPT-OSS 120B · monthly token cost
$27.34
per month
| Per user / month | $0.547 |
| Per session | $0.0061 |
| Per question | $0.0012 |
| Sessions / day | 150 |
| Tokens / month | 126.56M |