Inference Throughput

Measure AI inference throughput as generated tokens and completed requests per second.

How this calculator works

Enter generated tokens, completed requests, elapsed time (seconds). Select Calculate to apply the displayed formula and review the labeled results.

Formula / method

Tokens/second = generated tokens ÷ elapsed seconds; requests/second = requests ÷ elapsed seconds

Worked example

120,000 generated tokens across 1,000 requests in 600 seconds equals 200 tokens/s and 1.67 requests/s.

Assumptions and limitations

Counts cover the same measurement interval and represent completed work.

Averages can hide tail latency, batching effects, prompt-processing time, failed requests and hardware utilization.